Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
/* Copyright (C) 2007-2008 MySQL AB
|
|
|
|
|
|
|
|
This program is free software; you can redistribute it and/or modify
|
|
|
|
it under the terms of the GNU General Public License as published by
|
|
|
|
the Free Software Foundation; version 2 of the License.
|
|
|
|
|
|
|
|
This program is distributed in the hope that it will be useful,
|
|
|
|
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
GNU General Public License for more details.
|
|
|
|
|
|
|
|
You should have received a copy of the GNU General Public License
|
|
|
|
along with this program; if not, write to the Free Software
|
|
|
|
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
|
|
|
|
|
|
|
|
|
|
|
|
#include "mdl.h"
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
#include "debug_sync.h"
|
2009-12-02 17:31:57 +01:00
|
|
|
#include <hash.h>
|
|
|
|
#include <mysqld_error.h>
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
|
|
|
|
void notify_shared_lock(THD *thd, MDL_ticket *conflicting_ticket);
|
|
|
|
|
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
static bool mdl_initialized= 0;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
|
|
|
/**
|
|
|
|
A collection of all MDL locks. A singleton,
|
|
|
|
there is only one instance of the map in the server.
|
|
|
|
Maps MDL_key to MDL_lock instances.
|
|
|
|
*/
|
|
|
|
|
|
|
|
class MDL_map
|
|
|
|
{
|
|
|
|
public:
|
|
|
|
void init();
|
|
|
|
void destroy();
|
|
|
|
MDL_lock *find(const MDL_key *key);
|
|
|
|
MDL_lock *find_or_insert(const MDL_key *key);
|
|
|
|
void remove(MDL_lock *lock);
|
|
|
|
private:
|
|
|
|
bool move_from_hash_to_lock_mutex(MDL_lock *lock);
|
|
|
|
private:
|
|
|
|
/** All acquired locks in the server. */
|
|
|
|
HASH m_locks;
|
|
|
|
/* Protects access to m_locks hash. */
|
|
|
|
pthread_mutex_t m_mutex;
|
|
|
|
};
|
|
|
|
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
enum enum_deadlock_weight
|
|
|
|
{
|
|
|
|
MDL_DEADLOCK_WEIGHT_DML= 0,
|
|
|
|
MDL_DEADLOCK_WEIGHT_DDL= 100
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
A context of the recursive traversal through all contexts
|
|
|
|
in all sessions in search for deadlock.
|
|
|
|
*/
|
|
|
|
|
|
|
|
class Deadlock_detection_context
|
|
|
|
{
|
|
|
|
public:
|
|
|
|
Deadlock_detection_context(MDL_context *start_arg)
|
|
|
|
: start(start_arg),
|
|
|
|
victim(NULL),
|
|
|
|
current_search_depth(0)
|
|
|
|
{ }
|
|
|
|
MDL_context *start;
|
|
|
|
MDL_context *victim;
|
|
|
|
uint current_search_depth;
|
|
|
|
static const uint MAX_SEARCH_DEPTH= 1000;
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
Get a bit corresponding to enum_mdl_type value in a granted/waiting bitmaps
|
|
|
|
and compatibility matrices.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#define MDL_BIT(A) static_cast<MDL_lock::bitmap_t>(1U << A)
|
|
|
|
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
/**
|
2009-12-04 00:34:19 +01:00
|
|
|
The lock context. Created internally for an acquired lock.
|
2009-12-04 00:52:05 +01:00
|
|
|
For a given name, there exists only one MDL_lock instance,
|
2009-12-04 00:34:19 +01:00
|
|
|
and it exists only when the lock has been granted.
|
|
|
|
Can be seen as an MDL subsystem's version of TABLE_SHARE.
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
|
|
|
This is an abstract class which lacks information about
|
|
|
|
compatibility rules for lock types. They should be specified
|
|
|
|
in its descendants.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
*/
|
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
class MDL_lock
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
2009-12-04 00:52:05 +01:00
|
|
|
public:
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
typedef uchar bitmap_t;
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
class Ticket_list
|
|
|
|
{
|
|
|
|
public:
|
|
|
|
typedef I_P_List<MDL_ticket,
|
|
|
|
I_P_List_adapter<MDL_ticket,
|
|
|
|
&MDL_ticket::next_in_lock,
|
|
|
|
&MDL_ticket::prev_in_lock> >
|
|
|
|
List;
|
|
|
|
operator const List &() const { return m_list; }
|
|
|
|
Ticket_list() :m_bitmap(0) {}
|
|
|
|
|
|
|
|
void add_ticket(MDL_ticket *ticket);
|
|
|
|
void remove_ticket(MDL_ticket *ticket);
|
|
|
|
bool is_empty() const { return m_list.is_empty(); }
|
|
|
|
bitmap_t bitmap() const { return m_bitmap; }
|
|
|
|
private:
|
|
|
|
void clear_bit_if_not_in_list(enum_mdl_type type);
|
|
|
|
private:
|
|
|
|
/** List of tickets. */
|
|
|
|
List m_list;
|
|
|
|
/** Bitmap of types of tickets in this list. */
|
|
|
|
bitmap_t m_bitmap;
|
|
|
|
};
|
|
|
|
|
|
|
|
typedef Ticket_list::List::Iterator Ticket_iterator;
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
public:
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
/** The key of the object (data) being protected. */
|
2009-12-04 00:52:05 +01:00
|
|
|
MDL_key key;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
void *cached_object;
|
|
|
|
mdl_cached_object_release_hook cached_object_release_hook;
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
/**
|
|
|
|
Read-write lock protecting this lock context.
|
|
|
|
|
|
|
|
TODO/FIXME: Replace with RW-lock which will prefer readers
|
|
|
|
on all platforms and not only on Linux.
|
|
|
|
*/
|
|
|
|
rw_lock_t m_rwlock;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
bool is_empty() const
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
return (m_granted.is_empty() && m_waiting.is_empty());
|
|
|
|
}
|
|
|
|
|
|
|
|
virtual const bitmap_t *incompatible_granted_types_bitmap() const = 0;
|
|
|
|
virtual const bitmap_t *incompatible_waiting_types_bitmap() const = 0;
|
|
|
|
|
|
|
|
bool has_pending_conflicting_lock(enum_mdl_type type);
|
|
|
|
|
|
|
|
bool can_grant_lock(enum_mdl_type type, MDL_context *requstor_ctx) const;
|
|
|
|
|
|
|
|
inline static MDL_lock *create(const MDL_key *key);
|
|
|
|
|
|
|
|
void notify_shared_locks(MDL_context *ctx)
|
|
|
|
{
|
|
|
|
Ticket_iterator it(m_granted);
|
|
|
|
MDL_ticket *conflicting_ticket;
|
|
|
|
|
|
|
|
while ((conflicting_ticket= it++))
|
|
|
|
{
|
|
|
|
if (conflicting_ticket->get_ctx() != ctx)
|
|
|
|
notify_shared_lock(ctx->get_thd(), conflicting_ticket);
|
|
|
|
}
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
/**
|
|
|
|
Wake up contexts which are waiting to acquire lock on the object and
|
|
|
|
which may succeed now, when we released some lock on it or removed
|
|
|
|
some pending request from its waiters list (the latter can happen,
|
|
|
|
for example, when context trying to acquire exclusive on the object
|
|
|
|
lock is killed).
|
|
|
|
*/
|
|
|
|
void wake_up_waiters()
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
MDL_lock::Ticket_iterator it(m_waiting);
|
|
|
|
MDL_ticket *awake_ticket;
|
|
|
|
|
|
|
|
while ((awake_ticket= it++))
|
|
|
|
awake_ticket->get_ctx()->awake(MDL_context::NORMAL_WAKE_UP);
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
}
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
void remove_ticket(Ticket_list MDL_lock::*queue, MDL_ticket *ticket);
|
2009-12-04 00:52:05 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
bool find_deadlock(MDL_ticket *waiting_ticket,
|
|
|
|
Deadlock_detection_context *deadlock_ctx);
|
|
|
|
|
|
|
|
/** List of granted tickets for this lock. */
|
|
|
|
Ticket_list m_granted;
|
|
|
|
/** Tickets for contexts waiting to acquire a lock. */
|
|
|
|
Ticket_list m_waiting;
|
|
|
|
public:
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
MDL_lock(const MDL_key *key_arg)
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
: key(key_arg),
|
2009-12-04 00:52:05 +01:00
|
|
|
cached_object(NULL),
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
cached_object_release_hook(NULL),
|
|
|
|
m_ref_usage(0),
|
|
|
|
m_ref_release(0),
|
|
|
|
m_is_destroyed(FALSE)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
my_rwlock_init(&m_rwlock, NULL);
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
|
|
|
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
virtual ~MDL_lock()
|
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
rwlock_destroy(&m_rwlock);
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
}
|
|
|
|
inline static void destroy(MDL_lock *lock);
|
|
|
|
public:
|
|
|
|
/**
|
|
|
|
These three members are used to make it possible to separate
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
the mdl_locks.m_mutex mutex and MDL_lock::m_rwlock in
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
MDL_map::find_or_insert() for increased scalability.
|
|
|
|
The 'm_is_destroyed' member is only set by destroyers that
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
have both the mdl_locks.m_mutex and MDL_lock::m_rwlock, thus
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
holding any of the mutexes is sufficient to read it.
|
|
|
|
The 'm_ref_usage; is incremented under protection by
|
|
|
|
mdl_locks.m_mutex, but when 'm_is_destroyed' is set to TRUE, this
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
member is moved to be protected by the MDL_lock::m_rwlock.
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
This means that the MDL_map::find_or_insert() which only
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
holds the MDL_lock::m_rwlock can compare it to 'm_ref_release'
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
without acquiring mdl_locks.m_mutex again and if equal it can also
|
|
|
|
destroy the lock object safely.
|
|
|
|
The 'm_ref_release' is incremented under protection by
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
MDL_lock::m_rwlock.
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
Note since we are only interested in equality of these two
|
|
|
|
counters we don't have to worry about overflows as long as
|
|
|
|
their size is big enough to hold maximum number of concurrent
|
|
|
|
threads on the system.
|
|
|
|
*/
|
|
|
|
uint m_ref_usage;
|
|
|
|
uint m_ref_release;
|
|
|
|
bool m_is_destroyed;
|
|
|
|
};
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
|
|
|
|
2009-11-30 17:44:46 +01:00
|
|
|
/**
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
An implementation of the global metadata lock. The only locking modes
|
|
|
|
which are supported at the moment are SHARED and INTENTION EXCLUSIVE.
|
2009-11-30 17:44:46 +01:00
|
|
|
*/
|
|
|
|
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
class MDL_global_lock : public MDL_lock
|
2009-11-30 17:44:46 +01:00
|
|
|
{
|
2009-12-04 00:52:05 +01:00
|
|
|
public:
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
MDL_global_lock(const MDL_key *key_arg)
|
|
|
|
: MDL_lock(key_arg)
|
|
|
|
{ }
|
2009-12-04 00:52:05 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
virtual const bitmap_t *incompatible_granted_types_bitmap() const
|
|
|
|
{
|
|
|
|
return m_granted_incompatible;
|
|
|
|
}
|
|
|
|
virtual const bitmap_t *incompatible_waiting_types_bitmap() const
|
|
|
|
{
|
|
|
|
return m_waiting_incompatible;
|
|
|
|
}
|
|
|
|
|
|
|
|
private:
|
|
|
|
static const bitmap_t m_granted_incompatible[MDL_TYPE_END];
|
|
|
|
static const bitmap_t m_waiting_incompatible[MDL_TYPE_END];
|
2009-12-04 00:52:05 +01:00
|
|
|
};
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
|
|
|
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
/**
|
|
|
|
An implementation of a per-object lock. Supports SHARED, SHARED_UPGRADABLE,
|
|
|
|
SHARED HIGH PRIORITY and EXCLUSIVE locks.
|
|
|
|
*/
|
|
|
|
|
|
|
|
class MDL_object_lock : public MDL_lock
|
|
|
|
{
|
|
|
|
public:
|
|
|
|
MDL_object_lock(const MDL_key *key_arg)
|
|
|
|
: MDL_lock(key_arg)
|
|
|
|
{ }
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
virtual const bitmap_t *incompatible_granted_types_bitmap() const
|
|
|
|
{
|
|
|
|
return m_granted_incompatible;
|
|
|
|
}
|
|
|
|
virtual const bitmap_t *incompatible_waiting_types_bitmap() const
|
|
|
|
{
|
|
|
|
return m_waiting_incompatible;
|
|
|
|
}
|
|
|
|
|
|
|
|
private:
|
|
|
|
static const bitmap_t m_granted_incompatible[MDL_TYPE_END];
|
|
|
|
static const bitmap_t m_waiting_incompatible[MDL_TYPE_END];
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
};
|
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
static MDL_map mdl_locks;
|
2009-12-04 00:52:05 +01:00
|
|
|
|
|
|
|
extern "C"
|
|
|
|
{
|
|
|
|
static uchar *
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
mdl_locks_key(const uchar *record, size_t *length,
|
|
|
|
my_bool not_used __attribute__((unused)))
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
2009-12-04 00:52:05 +01:00
|
|
|
MDL_lock *lock=(MDL_lock*) record;
|
|
|
|
*length= lock->key.length();
|
|
|
|
return (uchar*) lock->key.ptr();
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
2009-12-04 00:52:05 +01:00
|
|
|
} /* extern "C" */
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
|
|
|
|
|
|
|
/**
|
2009-12-04 00:34:19 +01:00
|
|
|
Initialize the metadata locking subsystem.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
This function is called at server startup.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
In particular, initializes the new global mutex and
|
|
|
|
the associated condition variable: LOCK_mdl and COND_mdl.
|
|
|
|
These locking primitives are implementation details of the MDL
|
|
|
|
subsystem and are private to it.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
Note, that even though the new implementation adds acquisition
|
|
|
|
of a new global mutex to the execution flow of almost every SQL
|
|
|
|
statement, the design capitalizes on that to later save on
|
|
|
|
look ups in the table definition cache. This leads to reduced
|
|
|
|
contention overall and on LOCK_open in particular.
|
2009-12-04 00:52:05 +01:00
|
|
|
Please see the description of MDL_context::acquire_shared_lock()
|
|
|
|
for details.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
*/
|
|
|
|
|
|
|
|
void mdl_init()
|
|
|
|
{
|
2009-12-04 00:52:05 +01:00
|
|
|
DBUG_ASSERT(! mdl_initialized);
|
|
|
|
mdl_initialized= TRUE;
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
mdl_locks.init();
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
2009-12-04 00:34:19 +01:00
|
|
|
Release resources of metadata locking subsystem.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
Destroys the global mutex and the condition variable.
|
|
|
|
Called at server shutdown.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
*/
|
|
|
|
|
|
|
|
void mdl_destroy()
|
|
|
|
{
|
2009-12-04 00:52:05 +01:00
|
|
|
if (mdl_initialized)
|
|
|
|
{
|
|
|
|
mdl_initialized= FALSE;
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
mdl_locks.destroy();
|
2009-12-04 00:52:05 +01:00
|
|
|
}
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
/** Initialize the global hash containing all MDL locks. */
|
|
|
|
|
|
|
|
void MDL_map::init()
|
|
|
|
{
|
|
|
|
pthread_mutex_init(&m_mutex, NULL);
|
|
|
|
my_hash_init(&m_locks, &my_charset_bin, 16 /* FIXME */, 0, 0,
|
|
|
|
mdl_locks_key, 0, 0);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
/**
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
Destroy the global hash containing all MDL locks.
|
|
|
|
@pre It must be empty.
|
|
|
|
*/
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
void MDL_map::destroy()
|
|
|
|
{
|
|
|
|
DBUG_ASSERT(!m_locks.records);
|
|
|
|
pthread_mutex_destroy(&m_mutex);
|
|
|
|
my_hash_free(&m_locks);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
Find MDL_lock object corresponding to the key, create it
|
|
|
|
if it does not exist.
|
|
|
|
|
|
|
|
@retval non-NULL - Success. MDL_lock instance for the key with
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
locked MDL_lock::m_rwlock.
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
@retval NULL - Failure (OOM).
|
|
|
|
*/
|
|
|
|
|
|
|
|
MDL_lock* MDL_map::find_or_insert(const MDL_key *mdl_key)
|
|
|
|
{
|
|
|
|
MDL_lock *lock;
|
2010-02-04 07:25:32 +01:00
|
|
|
my_hash_value_type hash_value;
|
|
|
|
|
|
|
|
hash_value= my_calc_hash(&m_locks, mdl_key->ptr(), mdl_key->length());
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
|
|
|
retry:
|
|
|
|
pthread_mutex_lock(&m_mutex);
|
2010-02-04 07:25:32 +01:00
|
|
|
if (!(lock= (MDL_lock*) my_hash_search_using_hash_value(&m_locks,
|
|
|
|
hash_value,
|
|
|
|
mdl_key->ptr(),
|
|
|
|
mdl_key->length())))
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
{
|
|
|
|
lock= MDL_lock::create(mdl_key);
|
|
|
|
if (!lock || my_hash_insert(&m_locks, (uchar*)lock))
|
|
|
|
{
|
|
|
|
pthread_mutex_unlock(&m_mutex);
|
|
|
|
MDL_lock::destroy(lock);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (move_from_hash_to_lock_mutex(lock))
|
|
|
|
goto retry;
|
|
|
|
|
|
|
|
return lock;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
Find MDL_lock object corresponding to the key.
|
|
|
|
|
|
|
|
@retval non-NULL - MDL_lock instance for the key with locked
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
MDL_lock::m_rwlock.
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
@retval NULL - There was no MDL_lock for the key.
|
|
|
|
*/
|
|
|
|
|
|
|
|
MDL_lock* MDL_map::find(const MDL_key *mdl_key)
|
|
|
|
{
|
|
|
|
MDL_lock *lock;
|
2010-02-04 07:25:32 +01:00
|
|
|
my_hash_value_type hash_value;
|
|
|
|
|
|
|
|
hash_value= my_calc_hash(&m_locks, mdl_key->ptr(), mdl_key->length());
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
|
|
|
retry:
|
|
|
|
pthread_mutex_lock(&m_mutex);
|
2010-02-04 07:25:32 +01:00
|
|
|
if (!(lock= (MDL_lock*) my_hash_search_using_hash_value(&m_locks,
|
|
|
|
hash_value,
|
|
|
|
mdl_key->ptr(),
|
|
|
|
mdl_key->length())))
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
{
|
|
|
|
pthread_mutex_unlock(&m_mutex);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (move_from_hash_to_lock_mutex(lock))
|
|
|
|
goto retry;
|
|
|
|
|
|
|
|
return lock;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
Release mdl_locks.m_mutex mutex and lock MDL_lock::m_rwlock for lock
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
object from the hash. Handle situation when object was released
|
|
|
|
while the held no mutex.
|
|
|
|
|
|
|
|
@retval FALSE - Success.
|
|
|
|
@retval TRUE - Object was released while we held no mutex, caller
|
|
|
|
should re-try looking up MDL_lock object in the hash.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
*/
|
|
|
|
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
bool MDL_map::move_from_hash_to_lock_mutex(MDL_lock *lock)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
DBUG_ASSERT(! lock->m_is_destroyed);
|
|
|
|
safe_mutex_assert_owner(&m_mutex);
|
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
/*
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
We increment m_ref_usage which is a reference counter protected by
|
|
|
|
mdl_locks.m_mutex under the condition it is present in the hash and
|
|
|
|
m_is_destroyed is FALSE.
|
2009-12-04 00:52:05 +01:00
|
|
|
*/
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
lock->m_ref_usage++;
|
|
|
|
pthread_mutex_unlock(&m_mutex);
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
rw_wrlock(&lock->m_rwlock);
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
lock->m_ref_release++;
|
|
|
|
if (unlikely(lock->m_is_destroyed))
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
Object was released while we held no mutex, we need to
|
|
|
|
release it if no others hold references to it, while our own
|
|
|
|
reference count ensured that the object as such haven't got
|
|
|
|
its memory released yet. We can also safely compare
|
|
|
|
m_ref_usage and m_ref_release since the object is no longer
|
|
|
|
present in the hash so no one will be able to find it and
|
|
|
|
increment m_ref_usage anymore.
|
|
|
|
*/
|
|
|
|
uint ref_usage= lock->m_ref_usage;
|
|
|
|
uint ref_release= lock->m_ref_release;
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
rw_unlock(&lock->m_rwlock);
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
if (ref_usage == ref_release)
|
|
|
|
MDL_lock::destroy(lock);
|
|
|
|
return TRUE;
|
|
|
|
}
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
Destroy MDL_lock object or delegate this responsibility to
|
|
|
|
whatever thread that holds the last outstanding reference to
|
|
|
|
it.
|
|
|
|
*/
|
|
|
|
|
|
|
|
void MDL_map::remove(MDL_lock *lock)
|
|
|
|
{
|
|
|
|
uint ref_usage, ref_release;
|
|
|
|
|
|
|
|
if (lock->cached_object)
|
|
|
|
(*lock->cached_object_release_hook)(lock->cached_object);
|
|
|
|
|
|
|
|
/*
|
|
|
|
Destroy the MDL_lock object, but ensure that anyone that is
|
|
|
|
holding a reference to the object is not remaining, if so he
|
|
|
|
has the responsibility to release it.
|
|
|
|
|
|
|
|
Setting of m_is_destroyed to TRUE while holding _both_
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
mdl_locks.m_mutex and MDL_lock::m_rwlock mutexes transfers the
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
protection of m_ref_usage from mdl_locks.m_mutex to
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
MDL_lock::m_rwlock while removal of object from the hash makes
|
|
|
|
it read-only. Therefore whoever acquires MDL_lock::m_rwlock next
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
will see most up to date version of m_ref_usage.
|
|
|
|
|
|
|
|
This means that when m_is_destroyed is TRUE and we hold the
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
MDL_lock::m_rwlock we can safely read the m_ref_usage
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
member.
|
|
|
|
*/
|
|
|
|
pthread_mutex_lock(&m_mutex);
|
|
|
|
my_hash_delete(&m_locks, (uchar*) lock);
|
|
|
|
lock->m_is_destroyed= TRUE;
|
|
|
|
ref_usage= lock->m_ref_usage;
|
|
|
|
ref_release= lock->m_ref_release;
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
rw_unlock(&lock->m_rwlock);
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
pthread_mutex_unlock(&m_mutex);
|
|
|
|
if (ref_usage == ref_release)
|
|
|
|
MDL_lock::destroy(lock);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
Initialize a metadata locking context.
|
|
|
|
|
|
|
|
This is to be called when a new server connection is created.
|
|
|
|
*/
|
|
|
|
|
|
|
|
MDL_context::MDL_context()
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
:m_trans_sentinel(NULL),
|
|
|
|
m_thd(NULL),
|
|
|
|
m_needs_thr_lock_abort(FALSE),
|
|
|
|
m_waiting_for(NULL),
|
|
|
|
m_deadlock_weight(0),
|
|
|
|
m_signal(NO_WAKE_UP)
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
my_rwlock_init(&m_waiting_for_lock, NULL);
|
|
|
|
pthread_mutex_init(&m_signal_lock, NULL);
|
|
|
|
pthread_cond_init(&m_signal_cond, NULL);
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
2009-12-04 00:34:19 +01:00
|
|
|
Destroy metadata locking context.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
Assumes and asserts that there are no active or pending locks
|
|
|
|
associated with this context at the time of the destruction.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
Currently does nothing. Asserts that there are no pending
|
|
|
|
or satisfied lock requests. The pending locks must be released
|
|
|
|
prior to destruction. This is a new way to express the assertion
|
|
|
|
that all tables are closed before a connection is destroyed.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
*/
|
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
void MDL_context::destroy()
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
2009-12-04 00:52:05 +01:00
|
|
|
DBUG_ASSERT(m_tickets.is_empty());
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
|
|
|
|
rwlock_destroy(&m_waiting_for_lock);
|
|
|
|
pthread_mutex_destroy(&m_signal_lock);
|
|
|
|
pthread_cond_destroy(&m_signal_cond);
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
2009-12-04 00:34:19 +01:00
|
|
|
Initialize a lock request.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
This is to be used for every lock request.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
Note that initialization and allocation are split into two
|
|
|
|
calls. This is to allow flexible memory management of lock
|
|
|
|
requests. Normally a lock request is stored in statement memory
|
|
|
|
(e.g. is a member of struct TABLE_LIST), but we would also like
|
|
|
|
to allow allocation of lock requests in other memory roots,
|
|
|
|
for example in the grant subsystem, to lock privilege tables.
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
The MDL subsystem does not own or manage memory of lock requests.
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
|
2009-12-09 09:51:20 +01:00
|
|
|
@param mdl_namespace Id of namespace of object to be locked
|
|
|
|
@param db Name of database to which the object belongs
|
|
|
|
@param name Name of of the object
|
|
|
|
@param mdl_type The MDL lock type for the request.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
*/
|
|
|
|
|
2009-12-10 09:21:38 +01:00
|
|
|
void MDL_request::init(MDL_key::enum_mdl_namespace mdl_namespace,
|
2009-12-04 00:52:05 +01:00
|
|
|
const char *db_arg,
|
Backport of:
----------------------------------------------------------
revno: 2617.69.20
committer: Konstantin Osipov <kostja@sun.com>
branch nick: 5.4-4284-1-assert
timestamp: Thu 2009-08-13 18:29:55 +0400
message:
WL#4284 "Transactional DDL locking"
A review fix.
Since WL#4284 implementation separated MDL_request and MDL_ticket,
MDL_request becamse a utility object necessary only to get a ticket.
Store it by-value in TABLE_LIST with the intent to merge
MDL_request::key with table_list->table_name and table_list->db
in future.
Change the MDL subsystem to not require MDL_requests to
stay around till close_thread_tables().
Remove the list of requests from the MDL context.
Requests for shared metadata locks acquired in open_tables()
are only used as a list in recover_from_failed_open_table_attempt(),
which calls mdl_context.wait_for_locks() for this list.
To keep such list for recover_from_failed_open_table_attempt(),
introduce a context class (Open_table_context), that collects
all requests.
A lot of minor cleanups and simplications that became possible
with this change.
sql/event_db_repository.cc:
Remove alloc_mdl_requests(). Now MDL_request instance is a member
of TABLE_LIST, and init_one_table() initializes it.
sql/ha_ndbcluster_binlog.cc:
Remove now unnecessary declaration and initialization
of binlog_mdl_request.
sql/lock.cc:
No need to allocate MDL requests in lock_table_names() now.
sql/log.cc:
Use init_one_table() method, remove alloc_mdl_requests(),
which is now unnecessary.
sql/log_event.cc:
No need to allocate mdl_request separately now.
Use init_one_table() method.
sql/log_event_old.cc:
Update to the new signature of close_tables_for_reopen().
sql/mdl.cc:
Update try_acquire_exclusive_lock() to be more easy to use.
Function lock_table_name_if_not_cached() has been removed.
Make acquire_shared_lock() signature consistent with
try_acquire_exclusive_lock() signature.
Remove methods that are no longer used.
Update comments.
sql/mdl.h:
Implement an assignment operator that doesn't
copy MDL_key (MDL_key::operator= is private and
should remain private).
This is a hack to work-around assignment of TABLE_LIST
by value in several places. Such assignments violate
encapsulation, since only perform a shallow copy.
In most cases these assignments are a hack on their own.
sql/mysql_priv.h:
Update signatures of close_thread_tables() and close_tables_for_reopen().
sql/sp.cc:
Allocate TABLE_LIST in thd->mem_root.
Use init_one_table().
sql/sp_head.cc:
Use init_one_table().
Remove thd->locked_tables_root, it's no longer needed.
sql/sql_acl.cc:
Use init_mdl_requests() and init_one_table().
sql/sql_base.cc:
Update to new signatures of try_acquire_shared_lock() and
try_acquire_exclusive_lock().
Remove lock_table_name_if_not_cached().
Fix a bug in open_ltable() that would not return ER_LOCK_DEADLOCK
in case of a failed lock_tables() and a multi-statement
transaction.
Fix a bug in open_and_lock_tables_derived() that would
not return ER_LOCK_DEADLOCK in case of a multi-statement
transaction and a failure of lock_tables().
Move assignment of enum_open_table_action to a method of Open_table_context, a new class that maintains information
for backoff actions.
Minor rearrangements of the code.
Remove alloc_mdl_requests() in functions that work with system
tables: instead the patch ensures that callers always initialize
TABLE_LIST argument.
sql/sql_class.cc:
THD::locked_tables_root is no more.
sql/sql_class.h:
THD::locked_tables_root is no more.
Add a declaration for Open_table_context class.
sql/sql_delete.cc:
Update to use the simplified MDL API.
sql/sql_handler.cc:
TABLE_LIST::mdl_request is stored by-value now.
Ensure that mdl_request.ticket is NULL for every request
that is passed into MDL, to satisfy MDL asserts.
@ sql/sql_help.cc
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_help.cc:
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_insert.cc:
Remove assignment by-value of TABLE_LIST in
TABLEOP_HOOKS. We can't carry over a granted
MDL ticket from one table list to another.
sql/sql_parse.cc:
Change alloc_mdl_requests() -> init_mdl_requests().
@todo We can remove init_mdl_requests() altogether
in some places: all places that call add_table_to_list()
already have mdl requests initialized.
sql/sql_plugin.cc:
Use init_one_table().
THD::locked_tables_root is no more.
sql/sql_servers.cc:
Use init_one_table().
sql/sql_show.cc:
Update acquire_high_priority_shared_lock() to use
TABLE_LIST::mdl_request rather than allocate an own.
Fix get_trigger_table_impl() to use init_one_table(),
check for out of memory, follow the coding style.
sql/sql_table.cc:
Update to work with TABLE_LIST::mdl_request by-value.
Remove lock_table_name_if_not_cached().
The code that used to delegate to it is quite simple and
concise without it now.
sql/sql_udf.cc:
Use init_one_table().
sql/sql_update.cc:
Update to use the new signature of close_tables_for_reopen().
sql/table.cc:
Move re-setting of mdl_requests for prepared statements
and stored procedures from close_thread_tables() to
reinit_stmt_before_use().
Change alloc_mdl_requests() to init_mdl_requests().
init_mdl_requests() is a hack that can't be deleted
until we don't have a list-aware TABLE_LIST constructor.
Hopefully its use will be minimal
sql/table.h:
Change alloc_mdl_requests() to init_mdl_requests()
TABLE_LIST::mdl_request is stored by value.
sql/tztime.cc:
We no longer initialize mdl requests in open_system_tables_for*()
functions. Move this initialization closer to initialization
of the rest of TABLE_LIST members.
storage/myisammrg/ha_myisammrg.cc:
Simplify mdl_request initialization.
2009-12-08 10:57:07 +01:00
|
|
|
const char *name_arg,
|
|
|
|
enum enum_mdl_type mdl_type_arg)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
2009-12-09 09:51:20 +01:00
|
|
|
key.mdl_key_init(mdl_namespace, db_arg, name_arg);
|
Backport of:
----------------------------------------------------------
revno: 2617.69.20
committer: Konstantin Osipov <kostja@sun.com>
branch nick: 5.4-4284-1-assert
timestamp: Thu 2009-08-13 18:29:55 +0400
message:
WL#4284 "Transactional DDL locking"
A review fix.
Since WL#4284 implementation separated MDL_request and MDL_ticket,
MDL_request becamse a utility object necessary only to get a ticket.
Store it by-value in TABLE_LIST with the intent to merge
MDL_request::key with table_list->table_name and table_list->db
in future.
Change the MDL subsystem to not require MDL_requests to
stay around till close_thread_tables().
Remove the list of requests from the MDL context.
Requests for shared metadata locks acquired in open_tables()
are only used as a list in recover_from_failed_open_table_attempt(),
which calls mdl_context.wait_for_locks() for this list.
To keep such list for recover_from_failed_open_table_attempt(),
introduce a context class (Open_table_context), that collects
all requests.
A lot of minor cleanups and simplications that became possible
with this change.
sql/event_db_repository.cc:
Remove alloc_mdl_requests(). Now MDL_request instance is a member
of TABLE_LIST, and init_one_table() initializes it.
sql/ha_ndbcluster_binlog.cc:
Remove now unnecessary declaration and initialization
of binlog_mdl_request.
sql/lock.cc:
No need to allocate MDL requests in lock_table_names() now.
sql/log.cc:
Use init_one_table() method, remove alloc_mdl_requests(),
which is now unnecessary.
sql/log_event.cc:
No need to allocate mdl_request separately now.
Use init_one_table() method.
sql/log_event_old.cc:
Update to the new signature of close_tables_for_reopen().
sql/mdl.cc:
Update try_acquire_exclusive_lock() to be more easy to use.
Function lock_table_name_if_not_cached() has been removed.
Make acquire_shared_lock() signature consistent with
try_acquire_exclusive_lock() signature.
Remove methods that are no longer used.
Update comments.
sql/mdl.h:
Implement an assignment operator that doesn't
copy MDL_key (MDL_key::operator= is private and
should remain private).
This is a hack to work-around assignment of TABLE_LIST
by value in several places. Such assignments violate
encapsulation, since only perform a shallow copy.
In most cases these assignments are a hack on their own.
sql/mysql_priv.h:
Update signatures of close_thread_tables() and close_tables_for_reopen().
sql/sp.cc:
Allocate TABLE_LIST in thd->mem_root.
Use init_one_table().
sql/sp_head.cc:
Use init_one_table().
Remove thd->locked_tables_root, it's no longer needed.
sql/sql_acl.cc:
Use init_mdl_requests() and init_one_table().
sql/sql_base.cc:
Update to new signatures of try_acquire_shared_lock() and
try_acquire_exclusive_lock().
Remove lock_table_name_if_not_cached().
Fix a bug in open_ltable() that would not return ER_LOCK_DEADLOCK
in case of a failed lock_tables() and a multi-statement
transaction.
Fix a bug in open_and_lock_tables_derived() that would
not return ER_LOCK_DEADLOCK in case of a multi-statement
transaction and a failure of lock_tables().
Move assignment of enum_open_table_action to a method of Open_table_context, a new class that maintains information
for backoff actions.
Minor rearrangements of the code.
Remove alloc_mdl_requests() in functions that work with system
tables: instead the patch ensures that callers always initialize
TABLE_LIST argument.
sql/sql_class.cc:
THD::locked_tables_root is no more.
sql/sql_class.h:
THD::locked_tables_root is no more.
Add a declaration for Open_table_context class.
sql/sql_delete.cc:
Update to use the simplified MDL API.
sql/sql_handler.cc:
TABLE_LIST::mdl_request is stored by-value now.
Ensure that mdl_request.ticket is NULL for every request
that is passed into MDL, to satisfy MDL asserts.
@ sql/sql_help.cc
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_help.cc:
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_insert.cc:
Remove assignment by-value of TABLE_LIST in
TABLEOP_HOOKS. We can't carry over a granted
MDL ticket from one table list to another.
sql/sql_parse.cc:
Change alloc_mdl_requests() -> init_mdl_requests().
@todo We can remove init_mdl_requests() altogether
in some places: all places that call add_table_to_list()
already have mdl requests initialized.
sql/sql_plugin.cc:
Use init_one_table().
THD::locked_tables_root is no more.
sql/sql_servers.cc:
Use init_one_table().
sql/sql_show.cc:
Update acquire_high_priority_shared_lock() to use
TABLE_LIST::mdl_request rather than allocate an own.
Fix get_trigger_table_impl() to use init_one_table(),
check for out of memory, follow the coding style.
sql/sql_table.cc:
Update to work with TABLE_LIST::mdl_request by-value.
Remove lock_table_name_if_not_cached().
The code that used to delegate to it is quite simple and
concise without it now.
sql/sql_udf.cc:
Use init_one_table().
sql/sql_update.cc:
Update to use the new signature of close_tables_for_reopen().
sql/table.cc:
Move re-setting of mdl_requests for prepared statements
and stored procedures from close_thread_tables() to
reinit_stmt_before_use().
Change alloc_mdl_requests() to init_mdl_requests().
init_mdl_requests() is a hack that can't be deleted
until we don't have a list-aware TABLE_LIST constructor.
Hopefully its use will be minimal
sql/table.h:
Change alloc_mdl_requests() to init_mdl_requests()
TABLE_LIST::mdl_request is stored by value.
sql/tztime.cc:
We no longer initialize mdl requests in open_system_tables_for*()
functions. Move this initialization closer to initialization
of the rest of TABLE_LIST members.
storage/myisammrg/ha_myisammrg.cc:
Simplify mdl_request initialization.
2009-12-08 10:57:07 +01:00
|
|
|
type= mdl_type_arg;
|
2009-12-04 00:52:05 +01:00
|
|
|
ticket= NULL;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
2009-12-09 17:11:26 +01:00
|
|
|
/**
|
|
|
|
Initialize a lock request using pre-built MDL_key.
|
|
|
|
|
|
|
|
@sa MDL_request::init(namespace, db, name, type).
|
|
|
|
|
|
|
|
@param key_arg The pre-built MDL key for the request.
|
|
|
|
@param mdl_type_arg The MDL lock type for the request.
|
|
|
|
*/
|
|
|
|
|
|
|
|
void MDL_request::init(const MDL_key *key_arg,
|
|
|
|
enum enum_mdl_type mdl_type_arg)
|
|
|
|
{
|
|
|
|
key.mdl_key_init(key_arg);
|
|
|
|
type= mdl_type_arg;
|
|
|
|
ticket= NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
/**
|
2009-12-04 00:34:19 +01:00
|
|
|
Allocate and initialize one lock request.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
Same as mdl_init_lock(), but allocates the lock and the key buffer
|
|
|
|
on a memory root. Necessary to lock ad-hoc tables, e.g.
|
|
|
|
mysql.* tables of grant and data dictionary subsystems.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-09 09:51:20 +01:00
|
|
|
@param mdl_namespace Id of namespace of object to be locked
|
|
|
|
@param db Name of database to which object belongs
|
|
|
|
@param name Name of of object
|
|
|
|
@param root MEM_ROOT on which object should be allocated
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
@note The allocated lock request will have MDL_SHARED type.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
@retval 0 Error if out of memory
|
|
|
|
@retval non-0 Pointer to an object representing a lock request
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
*/
|
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
MDL_request *
|
2009-12-10 09:21:38 +01:00
|
|
|
MDL_request::create(MDL_key::enum_mdl_namespace mdl_namespace, const char *db,
|
Backport of:
----------------------------------------------------------
revno: 2617.69.20
committer: Konstantin Osipov <kostja@sun.com>
branch nick: 5.4-4284-1-assert
timestamp: Thu 2009-08-13 18:29:55 +0400
message:
WL#4284 "Transactional DDL locking"
A review fix.
Since WL#4284 implementation separated MDL_request and MDL_ticket,
MDL_request becamse a utility object necessary only to get a ticket.
Store it by-value in TABLE_LIST with the intent to merge
MDL_request::key with table_list->table_name and table_list->db
in future.
Change the MDL subsystem to not require MDL_requests to
stay around till close_thread_tables().
Remove the list of requests from the MDL context.
Requests for shared metadata locks acquired in open_tables()
are only used as a list in recover_from_failed_open_table_attempt(),
which calls mdl_context.wait_for_locks() for this list.
To keep such list for recover_from_failed_open_table_attempt(),
introduce a context class (Open_table_context), that collects
all requests.
A lot of minor cleanups and simplications that became possible
with this change.
sql/event_db_repository.cc:
Remove alloc_mdl_requests(). Now MDL_request instance is a member
of TABLE_LIST, and init_one_table() initializes it.
sql/ha_ndbcluster_binlog.cc:
Remove now unnecessary declaration and initialization
of binlog_mdl_request.
sql/lock.cc:
No need to allocate MDL requests in lock_table_names() now.
sql/log.cc:
Use init_one_table() method, remove alloc_mdl_requests(),
which is now unnecessary.
sql/log_event.cc:
No need to allocate mdl_request separately now.
Use init_one_table() method.
sql/log_event_old.cc:
Update to the new signature of close_tables_for_reopen().
sql/mdl.cc:
Update try_acquire_exclusive_lock() to be more easy to use.
Function lock_table_name_if_not_cached() has been removed.
Make acquire_shared_lock() signature consistent with
try_acquire_exclusive_lock() signature.
Remove methods that are no longer used.
Update comments.
sql/mdl.h:
Implement an assignment operator that doesn't
copy MDL_key (MDL_key::operator= is private and
should remain private).
This is a hack to work-around assignment of TABLE_LIST
by value in several places. Such assignments violate
encapsulation, since only perform a shallow copy.
In most cases these assignments are a hack on their own.
sql/mysql_priv.h:
Update signatures of close_thread_tables() and close_tables_for_reopen().
sql/sp.cc:
Allocate TABLE_LIST in thd->mem_root.
Use init_one_table().
sql/sp_head.cc:
Use init_one_table().
Remove thd->locked_tables_root, it's no longer needed.
sql/sql_acl.cc:
Use init_mdl_requests() and init_one_table().
sql/sql_base.cc:
Update to new signatures of try_acquire_shared_lock() and
try_acquire_exclusive_lock().
Remove lock_table_name_if_not_cached().
Fix a bug in open_ltable() that would not return ER_LOCK_DEADLOCK
in case of a failed lock_tables() and a multi-statement
transaction.
Fix a bug in open_and_lock_tables_derived() that would
not return ER_LOCK_DEADLOCK in case of a multi-statement
transaction and a failure of lock_tables().
Move assignment of enum_open_table_action to a method of Open_table_context, a new class that maintains information
for backoff actions.
Minor rearrangements of the code.
Remove alloc_mdl_requests() in functions that work with system
tables: instead the patch ensures that callers always initialize
TABLE_LIST argument.
sql/sql_class.cc:
THD::locked_tables_root is no more.
sql/sql_class.h:
THD::locked_tables_root is no more.
Add a declaration for Open_table_context class.
sql/sql_delete.cc:
Update to use the simplified MDL API.
sql/sql_handler.cc:
TABLE_LIST::mdl_request is stored by-value now.
Ensure that mdl_request.ticket is NULL for every request
that is passed into MDL, to satisfy MDL asserts.
@ sql/sql_help.cc
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_help.cc:
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_insert.cc:
Remove assignment by-value of TABLE_LIST in
TABLEOP_HOOKS. We can't carry over a granted
MDL ticket from one table list to another.
sql/sql_parse.cc:
Change alloc_mdl_requests() -> init_mdl_requests().
@todo We can remove init_mdl_requests() altogether
in some places: all places that call add_table_to_list()
already have mdl requests initialized.
sql/sql_plugin.cc:
Use init_one_table().
THD::locked_tables_root is no more.
sql/sql_servers.cc:
Use init_one_table().
sql/sql_show.cc:
Update acquire_high_priority_shared_lock() to use
TABLE_LIST::mdl_request rather than allocate an own.
Fix get_trigger_table_impl() to use init_one_table(),
check for out of memory, follow the coding style.
sql/sql_table.cc:
Update to work with TABLE_LIST::mdl_request by-value.
Remove lock_table_name_if_not_cached().
The code that used to delegate to it is quite simple and
concise without it now.
sql/sql_udf.cc:
Use init_one_table().
sql/sql_update.cc:
Update to use the new signature of close_tables_for_reopen().
sql/table.cc:
Move re-setting of mdl_requests for prepared statements
and stored procedures from close_thread_tables() to
reinit_stmt_before_use().
Change alloc_mdl_requests() to init_mdl_requests().
init_mdl_requests() is a hack that can't be deleted
until we don't have a list-aware TABLE_LIST constructor.
Hopefully its use will be minimal
sql/table.h:
Change alloc_mdl_requests() to init_mdl_requests()
TABLE_LIST::mdl_request is stored by value.
sql/tztime.cc:
We no longer initialize mdl requests in open_system_tables_for*()
functions. Move this initialization closer to initialization
of the rest of TABLE_LIST members.
storage/myisammrg/ha_myisammrg.cc:
Simplify mdl_request initialization.
2009-12-08 10:57:07 +01:00
|
|
|
const char *name, enum_mdl_type mdl_type,
|
|
|
|
MEM_ROOT *root)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
2009-12-04 00:52:05 +01:00
|
|
|
MDL_request *mdl_request;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
if (!(mdl_request= (MDL_request*) alloc_root(root, sizeof(MDL_request))))
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
return NULL;
|
|
|
|
|
2009-12-09 09:51:20 +01:00
|
|
|
mdl_request->init(mdl_namespace, db, name, mdl_type);
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
return mdl_request;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
uint MDL_request::get_deadlock_weight() const
|
|
|
|
{
|
|
|
|
return key.mdl_namespace() == MDL_key::GLOBAL ||
|
|
|
|
type > MDL_SHARED_NO_WRITE ?
|
|
|
|
MDL_DEADLOCK_WEIGHT_DDL : MDL_DEADLOCK_WEIGHT_DML;
|
|
|
|
}
|
|
|
|
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
/**
|
2009-12-04 00:52:05 +01:00
|
|
|
Auxiliary functions needed for creation/destruction of MDL_lock objects.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
@note Also chooses an MDL_lock descendant appropriate for object namespace.
|
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
@todo This naive implementation should be replaced with one that saves
|
|
|
|
on memory allocation by reusing released objects.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
*/
|
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
inline MDL_lock *MDL_lock::create(const MDL_key *mdl_key)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
switch (mdl_key->mdl_namespace())
|
|
|
|
{
|
|
|
|
case MDL_key::GLOBAL:
|
|
|
|
return new MDL_global_lock(mdl_key);
|
|
|
|
default:
|
|
|
|
return new MDL_object_lock(mdl_key);
|
|
|
|
}
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
void MDL_lock::destroy(MDL_lock *lock)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
|
|
|
delete lock;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
/**
|
2009-12-04 00:52:05 +01:00
|
|
|
Auxiliary functions needed for creation/destruction of MDL_ticket
|
2009-12-04 00:34:19 +01:00
|
|
|
objects.
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
@todo This naive implementation should be replaced with one that saves
|
|
|
|
on memory allocation by reusing released objects.
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
*/
|
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
MDL_ticket *MDL_ticket::create(MDL_context *ctx_arg, enum_mdl_type type_arg)
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
{
|
2009-12-04 00:52:05 +01:00
|
|
|
return new MDL_ticket(ctx_arg, type_arg);
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
void MDL_ticket::destroy(MDL_ticket *ticket)
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
{
|
|
|
|
delete ticket;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2009-12-02 17:31:57 +01:00
|
|
|
/**
|
2009-12-04 00:34:19 +01:00
|
|
|
Helper functions and macros to be used for killable waiting in metadata
|
|
|
|
locking subsystem.
|
2009-12-02 17:31:57 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
@sa THD::enter_cond()/exit_cond()/killed.
|
2009-12-02 17:31:57 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
@note We can't use THD::enter_cond()/exit_cond()/killed directly here
|
Backport of revno ## 2617.31.1, 2617.31.3, 2617.31.4, 2617.31.5,
2617.31.12, 2617.31.15, 2617.31.15, 2617.31.16, 2617.43.1
- initial changeset that introduced the fix for
Bug#989 and follow up fixes for all test suite failures
introduced in the initial changeset.
------------------------------------------------------------
revno: 2617.31.1
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Fri 2009-03-06 19:17:00 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
Currently the MySQL server does not keep metadata locks on
schema objects for the duration of a transaction, thus failing
to guarantee the integrity of the schema objects being used
during the transaction and to protect then from concurrent
DDL operations. This also poses a problem for replication as
a DDL operation might be replicated even thought there are
active transactions using the object being modified.
The solution is to defer the release of metadata locks until
a active transaction is either committed or rolled back. This
prevents other statements from modifying the table for the
entire duration of the transaction. This provides commitment
ordering for guaranteeing serializability across multiple
transactions.
- Incompatible change:
If MySQL's metadata locking system encounters a lock conflict,
the usual schema is to use the try and back-off technique to
avoid deadlocks -- this schema consists in releasing all locks
and trying to acquire them all in one go.
But in a transactional context this algorithm can't be utilized
as its not possible to release locks acquired during the course
of the transaction without breaking the transaction commitments.
To avoid deadlocks in this case, the ER_LOCK_DEADLOCK will be
returned if a lock conflict is encountered during a transaction.
Let's consider an example:
A transaction has two statements that modify table t1, then table
t2, and then commits. The first statement of the transaction will
acquire a shared metadata lock on table t1, and it will be kept
utill COMMIT to ensure serializability.
At the moment when the second statement attempts to acquire a
shared metadata lock on t2, a concurrent ALTER or DROP statement
might have locked t2 exclusively. The prescription of the current
locking protocol is that the acquirer of the shared lock backs off
-- gives up all his current locks and retries. This implies that
the entire multi-statement transaction has to be rolled back.
- Incompatible change:
FLUSH commands such as FLUSH PRIVILEGES and FLUSH TABLES WITH READ
LOCK won't cause locked tables to be implicitly unlocked anymore.
mysql-test/extra/binlog_tests/drop_table.test:
Add test case for Bug#989.
mysql-test/extra/binlog_tests/mix_innodb_myisam_binlog.test:
Fix test case to reflect the fact that transactions now hold
metadata locks for the duration of a transaction.
mysql-test/include/mix1.inc:
Fix test case to reflect the fact that transactions now hold
metadata locks for the duration of a transaction.
mysql-test/include/mix2.inc:
Fix test case to reflect the fact that transactions now hold
metadata locks for the duration of a transaction.
mysql-test/r/flush_block_commit.result:
Update test case result (WL#4284).
mysql-test/r/flush_block_commit_notembedded.result:
Update test case result (WL#4284).
mysql-test/r/innodb.result:
Update test case result (WL#4284).
mysql-test/r/innodb_mysql.result:
Update test case result (WL#4284).
mysql-test/r/lock.result:
Add test case result for an effect of WL#4284/Bug#989
(all locks should be released when a connection terminates).
mysql-test/r/mix2_myisam.result:
Update test case result (effects of WL#4284/Bug#989).
mysql-test/r/not_embedded_server.result:
Update test case result (effects of WL#4284/Bug#989).
Add a test case for interaction of WL#4284 and FLUSH PRIVILEGES.
mysql-test/r/partition_innodb_semi_consistent.result:
Update test case result (effects of WL#4284/Bug#989).
mysql-test/r/partition_sync.result:
Temporarily disable the test case for Bug#43867,
which will be fixed by a subsequent backport.
mysql-test/r/ps.result:
Add a test case for effect of PREPARE on transactional
locks: we take a savepoint at beginning of PREAPRE
and release it at the end. Thus PREPARE does not
accumulate metadata locks (Bug#989/WL#4284).
mysql-test/r/read_only_innodb.result:
Update test case result (effects of WL#4284/Bug#989).
mysql-test/suite/binlog/r/binlog_row_drop_tbl.result:
Add a test case result (WL#4284/Bug#989).
mysql-test/suite/binlog/r/binlog_row_mix_innodb_myisam.result:
Update test case result (effects of WL#4284/Bug#989).
mysql-test/suite/binlog/r/binlog_stm_drop_tbl.result:
Add a test case result (WL#4284/Bug#989).
mysql-test/suite/binlog/r/binlog_stm_mix_innodb_myisam.result:
Update test case result (effects of WL#4284/Bug#989).
mysql-test/suite/binlog/r/binlog_unsafe.result:
A side effect of Bug#989 -- slightly different table map ids.
mysql-test/suite/binlog/t/binlog_row_drop_tbl.test:
Add a test case for WL#4284/Bug#989.
mysql-test/suite/binlog/t/binlog_stm_drop_tbl.test:
Add a test case for WL#4284/Bug#989.
mysql-test/suite/binlog/t/binlog_stm_row.test:
Update to the new state name. This
is actually a follow up to another patch for WL#4284,
that changes Locked thread state to Table lock.
mysql-test/suite/ndb/r/ndb_index_ordered.result:
Remove result for disabled part of the test case.
mysql-test/suite/ndb/t/disabled.def:
Temporarily disable a test case (Bug#45621).
mysql-test/suite/ndb/t/ndb_index_ordered.test:
Disable a part of a test case (needs update to
reflect semantics of Bug#989).
mysql-test/suite/rpl/t/disabled.def:
Disable tests made meaningless by transactional metadata
locking.
mysql-test/suite/sys_vars/r/autocommit_func.result:
Add a commit (Bug#989).
mysql-test/suite/sys_vars/t/autocommit_func.test:
Add a commit (Bug#989).
mysql-test/t/flush_block_commit.test:
Fix test case to reflect the fact that transactions now hold
metadata locks for the duration of a transaction.
mysql-test/t/flush_block_commit_notembedded.test:
Fix test case to reflect the fact that transactions now hold
metadata locks for the duration of a transaction.
Add a test case for transaction-scope locks and the global
read lock (Bug#989/WL#4284).
mysql-test/t/innodb.test:
Fix test case to reflect the fact that transactions now hold
metadata locks for the duration of a transaction
(effects of Bug#989/WL#4284).
mysql-test/t/lock.test:
Add a test case for Bug#989/WL#4284.
mysql-test/t/not_embedded_server.test:
Add a test case for Bug#989/WL#4284.
mysql-test/t/partition_innodb_semi_consistent.test:
Replace TRUNCATE with DELETE, to not issue
an implicit commit of a transaction, and not depend
on metadata locks.
mysql-test/t/partition_sync.test:
Temporarily disable the test case for Bug#43867,
which needs a fix to be backported from 6.0.
mysql-test/t/ps.test:
Add a test case for semantics of PREPARE and transaction-scope
locks: metadata locks on tables used in PREPARE are enclosed into a
temporary savepoint, taken at the beginning of PREPARE,
and released at the end. Thus PREPARE does not effect
what locks a transaction owns.
mysql-test/t/read_only_innodb.test:
Fix test case to reflect the fact that transactions now hold
metadata locks for the duration of a transaction
(Bug#989/WL#4284).
Wait for the read_only statement to actually flush tables before
sending other concurrent statements that depend on its state.
mysql-test/t/xa.test:
Fix test case to reflect the fact that transactions now hold
metadata locks for the duration of a transaction
(Bug#989/WL#4284).
sql/ha_ndbcluster_binlog.cc:
Backport bits of changes of ha_ndbcluster_binlog.cc
from 6.0, to fix the failing binlog test suite with
WL#4284. WL#4284 implementation does not work
with 5.1 implementation of ndbcluster binlog index.
sql/log_event.cc:
Release metadata locks after issuing a commit.
sql/mdl.cc:
Style changes (WL#4284).
sql/mysql_priv.h:
Rename parameter to match the name used in the definition (WL#4284).
sql/rpl_injector.cc:
Release metadata locks on commit (WL#4284).
sql/rpl_rli.cc:
Remove assert made meaningless, metadata locks are released
at the end of the transaction.
sql/set_var.cc:
Close tables and release locks if autocommit mode is set.
sql/slave.cc:
Release metadata locks after a rollback.
sql/sql_acl.cc:
Don't implicitly unlock locked tables. Issue a implicit commit
at the end and unlock tables.
sql/sql_base.cc:
Defer the release of metadata locks when closing tables
if not required to.
Issue a deadlock error if the locking protocol requires
that a transaction re-acquire its locks.
Release metadata locks when closing tables for reopen.
sql/sql_class.cc:
Release metadata locks if the thread is killed.
sql/sql_parse.cc:
Release metadata locks after implicitly committing a active
transaction, or after explicit commits or rollbacks.
sql/sql_plugin.cc:
Allocate MDL request on the stack as the use of the table
is contained within the function. It will be removed from
the context once close_thread_tables is called at the end
of the function.
sql/sql_prepare.cc:
The problem is that the prepare phase of the CREATE TABLE
statement takes a exclusive metadata lock lock and this can
cause a self-deadlock the thread already holds a shared lock
on the table being that should be created.
The solution is to make the prepare phase take a shared
metadata lock when preparing a CREATE TABLE statement. The
execution of the statement will still acquire a exclusive
lock, but won't cause any problem as it issues a implicit
commit.
After some discussions with stakeholders it has been decided that
metadata locks acquired during a PREPARE statement must be released
once the statement is prepared even if it is prepared within a multi
statement transaction.
sql/sql_servers.cc:
Don't implicitly unlock locked tables. Issue a implicit commit
at the end and unlock tables.
sql/sql_table.cc:
Close table and release metadata locks after a admin operation.
sql/table.h:
The problem is that the prepare phase of the CREATE TABLE
statement takes a exclusive metadata lock lock and this can
cause a self-deadlock the thread already holds a shared lock
on the table being that should be created.
The solution is to make the prepare phase take a shared
metadata lock when preparing a CREATE TABLE statement. The
execution of the statement will still acquire a exclusive
lock, but won't cause any problem as it issues a implicit
commit.
sql/transaction.cc:
Release metadata locks after the implicitly committed due
to a new transaction being started. Also, release metadata
locks acquired after a savepoint if the transaction is rolled
back to the save point.
The problem is that in some cases transaction-long metadata locks
could be released before the transaction was committed. This could
happen when a active transaction was ended by a "START TRANSACTION"
or "BEGIN" statement, in which case the metadata locks would be
released before the actual commit of the active transaction.
The solution is to defer the release of metadata locks to after the
transaction has been implicitly committed. No test case is provided
as the effort to provide one is too disproportional to the size of
the fix.
2009-12-05 00:02:48 +01:00
|
|
|
since this will make metadata subsystem dependent on THD class
|
2009-12-04 00:34:19 +01:00
|
|
|
and thus prevent us from writing unit tests for it. And usage of
|
|
|
|
wrapper functions to access THD::killed/enter_cond()/exit_cond()
|
|
|
|
will probably introduce too much overhead.
|
2009-12-02 17:31:57 +01:00
|
|
|
*/
|
|
|
|
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
#define MDL_ENTER_COND(A, B, C, D) \
|
|
|
|
mdl_enter_cond(A, B, C, D, __func__, __FILE__, __LINE__)
|
2009-12-02 17:31:57 +01:00
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
static inline const char *mdl_enter_cond(THD *thd,
|
2009-12-02 17:31:57 +01:00
|
|
|
st_my_thread_var *mysys_var,
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
pthread_cond_t *cond,
|
|
|
|
pthread_mutex_t *mutex,
|
2009-12-02 17:31:57 +01:00
|
|
|
const char *calling_func,
|
|
|
|
const char *calling_file,
|
|
|
|
const unsigned int calling_line)
|
|
|
|
{
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
safe_mutex_assert_owner(mutex);
|
2009-12-02 17:31:57 +01:00
|
|
|
|
2010-02-02 00:22:16 +01:00
|
|
|
mysys_var->current_mutex= (mysql_mutex_t*) mutex;
|
|
|
|
mysys_var->current_cond= (mysql_cond_t*) cond;
|
2009-12-02 17:31:57 +01:00
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
DEBUG_SYNC(thd, "mdl_enter_cond");
|
|
|
|
|
|
|
|
return set_thd_proc_info(thd, "Waiting for table",
|
2009-12-02 17:31:57 +01:00
|
|
|
calling_func, calling_file, calling_line);
|
|
|
|
}
|
|
|
|
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
#define MDL_EXIT_COND(A, B, C, D) \
|
|
|
|
mdl_exit_cond(A, B, C, D, __func__, __FILE__, __LINE__)
|
2009-12-02 17:31:57 +01:00
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
static inline void mdl_exit_cond(THD *thd,
|
2009-12-02 17:31:57 +01:00
|
|
|
st_my_thread_var *mysys_var,
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
pthread_mutex_t *mutex,
|
2009-12-02 17:31:57 +01:00
|
|
|
const char* old_msg,
|
|
|
|
const char *calling_func,
|
|
|
|
const char *calling_file,
|
|
|
|
const unsigned int calling_line)
|
|
|
|
{
|
2010-02-02 00:22:16 +01:00
|
|
|
DBUG_ASSERT(mutex == (pthread_mutex_t*) mysys_var->current_mutex);
|
2009-12-02 17:31:57 +01:00
|
|
|
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
pthread_mutex_unlock(mutex);
|
2010-02-02 00:22:16 +01:00
|
|
|
mysql_mutex_lock(&mysys_var->mutex);
|
2009-12-02 17:31:57 +01:00
|
|
|
mysys_var->current_mutex= 0;
|
|
|
|
mysys_var->current_cond= 0;
|
2010-02-02 00:22:16 +01:00
|
|
|
mysql_mutex_unlock(&mysys_var->mutex);
|
2009-12-02 17:31:57 +01:00
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
DEBUG_SYNC(thd, "mdl_exit_cond");
|
|
|
|
|
|
|
|
(void) set_thd_proc_info(thd, old_msg, calling_func,
|
2009-12-02 17:31:57 +01:00
|
|
|
calling_file, calling_line);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
MDL_context::mdl_signal_type MDL_context::wait()
|
|
|
|
{
|
|
|
|
const char *old_msg;
|
|
|
|
st_my_thread_var *mysys_var= my_thread_var;
|
|
|
|
mdl_signal_type result;
|
2009-12-04 00:34:19 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
pthread_mutex_lock(&m_signal_lock);
|
2009-12-04 00:34:19 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
old_msg= MDL_ENTER_COND(m_thd, mysys_var, &m_signal_cond, &m_signal_lock);
|
2009-12-01 14:55:45 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
while (! m_signal && !mysys_var->abort)
|
|
|
|
pthread_cond_wait(&m_signal_cond, &m_signal_lock);
|
|
|
|
|
|
|
|
result= m_signal;
|
|
|
|
|
|
|
|
MDL_EXIT_COND(m_thd, mysys_var, &m_signal_lock, old_msg);
|
|
|
|
|
|
|
|
return result;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
MDL_context::mdl_signal_type MDL_context::timed_wait(ulong timeout)
|
2009-12-01 14:55:45 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
struct timespec abstime;
|
|
|
|
const char *old_msg;
|
|
|
|
mdl_signal_type result;
|
|
|
|
st_my_thread_var *mysys_var= my_thread_var;
|
|
|
|
|
|
|
|
pthread_mutex_lock(&m_signal_lock);
|
|
|
|
|
|
|
|
old_msg= MDL_ENTER_COND(m_thd, mysys_var, &m_signal_cond, &m_signal_lock);
|
|
|
|
|
|
|
|
if (! m_signal)
|
2009-12-01 14:55:45 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
set_timespec(abstime, timeout);
|
|
|
|
pthread_cond_timedwait(&m_signal_cond, &m_signal_lock, &abstime);
|
2009-12-01 14:55:45 +01:00
|
|
|
}
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
|
|
|
|
result= (m_signal != NO_WAKE_UP) ? m_signal : TIMEOUT_WAKE_UP;
|
|
|
|
|
|
|
|
MDL_EXIT_COND(m_thd, mysys_var, &m_signal_lock, old_msg);
|
|
|
|
|
|
|
|
return result;
|
2009-12-01 14:55:45 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
Clear bit corresponding to the type of metadata lock in bitmap representing
|
|
|
|
set of such types if list of tickets does not contain ticket with such type.
|
|
|
|
|
|
|
|
@param[in,out] bitmap Bitmap representing set of types of locks.
|
|
|
|
@param[in] list List to inspect.
|
|
|
|
@param[in] type Type of metadata lock to look up in the list.
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
*/
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
void MDL_lock::Ticket_list::clear_bit_if_not_in_list(enum_mdl_type type)
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
MDL_lock::Ticket_iterator it(m_list);
|
|
|
|
const MDL_ticket *ticket;
|
|
|
|
|
|
|
|
while ((ticket= it++))
|
|
|
|
if (ticket->get_type() == type)
|
|
|
|
return;
|
|
|
|
m_bitmap&= ~ MDL_BIT(type);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
Add ticket to MDL_lock's list of waiting requests and
|
|
|
|
update corresponding bitmap of lock types.
|
|
|
|
*/
|
|
|
|
|
|
|
|
void MDL_lock::Ticket_list::add_ticket(MDL_ticket *ticket)
|
|
|
|
{
|
2010-02-01 15:38:50 +01:00
|
|
|
/*
|
|
|
|
Ticket being added to the list must have MDL_ticket::m_lock set,
|
|
|
|
since for such tickets methods accessing this member might be
|
|
|
|
called by other threads.
|
|
|
|
*/
|
|
|
|
DBUG_ASSERT(ticket->get_lock());
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
m_list.push_front(ticket);
|
|
|
|
m_bitmap|= MDL_BIT(ticket->get_type());
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
Remove ticket from MDL_lock's list of requests and
|
|
|
|
update corresponding bitmap of lock types.
|
|
|
|
*/
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
void MDL_lock::Ticket_list::remove_ticket(MDL_ticket *ticket)
|
|
|
|
{
|
|
|
|
m_list.remove(ticket);
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
/*
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
Check if waiting queue has another ticket with the same type as
|
|
|
|
one which was removed. If there is no such ticket, i.e. we have
|
|
|
|
removed last ticket of particular type, then we need to update
|
|
|
|
bitmap of waiting ticket's types.
|
|
|
|
Note that in most common case, i.e. when shared lock is removed
|
|
|
|
from waiting queue, we are likely to find ticket of the same
|
|
|
|
type early without performing full iteration through the list.
|
|
|
|
So this method should not be too expensive.
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
*/
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
clear_bit_if_not_in_list(ticket->get_type());
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
Compatibility (or rather "incompatibility") matrices for global metadata
|
|
|
|
lock. Arrays of bitmaps which elements specify which granted/waiting locks
|
|
|
|
are incompatible with type of lock being requested.
|
|
|
|
|
|
|
|
Here is how types of individual locks are translated to type of global lock:
|
|
|
|
|
|
|
|
----------------+-------------+
|
|
|
|
Type of request | Correspond. |
|
|
|
|
for indiv. lock | global lock |
|
|
|
|
----------------+-------------+
|
|
|
|
S, SH, SR, SW | IS |
|
|
|
|
SNW, SNRW, X | IX |
|
|
|
|
SNW, SNRW -> X | IX (*) |
|
|
|
|
|
|
|
|
The first array specifies if particular type of request can be satisfied
|
|
|
|
if there is granted global lock of certain type.
|
|
|
|
|
|
|
|
| Type of active |
|
|
|
|
Request | global lock |
|
|
|
|
type | IS(**) IX S |
|
|
|
|
---------+----------------+
|
|
|
|
IS | + + + |
|
|
|
|
IX | + + - |
|
|
|
|
S | + - + |
|
|
|
|
|
|
|
|
The second array specifies if particular type of request can be satisfied
|
|
|
|
if there is already waiting request for the global lock of certain type.
|
|
|
|
I.e. it specifies what is the priority of different lock types.
|
|
|
|
|
|
|
|
| Pending |
|
|
|
|
Request | global lock |
|
|
|
|
type | IS(**) IX S |
|
|
|
|
---------+--------------+
|
|
|
|
IS | + + + |
|
|
|
|
IX | + + - |
|
|
|
|
S | + + + |
|
2009-12-01 14:55:45 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
Here: "+" -- means that request can be satisfied
|
|
|
|
"-" -- means that request can't be satisfied and should wait
|
2009-12-01 14:55:45 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
(*) Since for upgradable locks we always take intention exclusive global
|
|
|
|
lock at the same time when obtaining the shared lock, there is no
|
|
|
|
need to obtain such lock during the upgrade itself.
|
|
|
|
(**) Since intention shared global locks are compatible with all other
|
|
|
|
type of locks we don't even have any accounting for them.
|
|
|
|
*/
|
|
|
|
|
|
|
|
const MDL_lock::bitmap_t MDL_global_lock::m_granted_incompatible[MDL_TYPE_END] =
|
|
|
|
{
|
|
|
|
MDL_BIT(MDL_SHARED), MDL_BIT(MDL_INTENTION_EXCLUSIVE), 0, 0, 0, 0, 0, 0
|
|
|
|
};
|
|
|
|
|
|
|
|
const MDL_lock::bitmap_t MDL_global_lock::m_waiting_incompatible[MDL_TYPE_END] =
|
|
|
|
{
|
|
|
|
MDL_BIT(MDL_SHARED), 0, 0, 0, 0, 0, 0, 0
|
|
|
|
};
|
2009-12-01 14:55:45 +01:00
|
|
|
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
/**
|
|
|
|
Compatibility (or rather "incompatibility") matrices for per-object
|
|
|
|
metadata lock. Arrays of bitmaps which elements specify which granted/
|
|
|
|
waiting locks are incompatible with type of lock being requested.
|
|
|
|
|
|
|
|
The first array specifies if particular type of request can be satisfied
|
|
|
|
if there is granted lock of certain type.
|
|
|
|
|
|
|
|
Request | Granted requests for lock |
|
|
|
|
type | S SH SR SW SNW SNRW X |
|
|
|
|
----------+------------------------------+
|
|
|
|
S | + + + + + + - |
|
|
|
|
SH | + + + + + + - |
|
|
|
|
SR | + + + + + - - |
|
|
|
|
SW | + + + + - - - |
|
|
|
|
SNW | + + + - - - - |
|
|
|
|
SNRW | + + - - - - - |
|
|
|
|
X | - - - - - - - |
|
|
|
|
SNW -> X | - - - 0 0 0 0 |
|
|
|
|
SNRW -> X | - - 0 0 0 0 0 |
|
|
|
|
|
|
|
|
The second array specifies if particular type of request can be satisfied
|
|
|
|
if there is waiting request for the same lock of certain type. In other
|
|
|
|
words it specifies what is the priority of different lock types.
|
|
|
|
|
|
|
|
Request | Pending requests for lock |
|
|
|
|
type | S SH SR SW SNW SNRW X |
|
|
|
|
----------+-----------------------------+
|
|
|
|
S | + + + + + + - |
|
|
|
|
SH | + + + + + + + |
|
|
|
|
SR | + + + + + - - |
|
|
|
|
SW | + + + + - - - |
|
|
|
|
SNW | + + + + + + - |
|
|
|
|
SNRW | + + + + + + - |
|
|
|
|
X | + + + + + + + |
|
|
|
|
SNW -> X | + + + + + + + |
|
|
|
|
SNRW -> X | + + + + + + + |
|
2009-12-01 14:55:45 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
Here: "+" -- means that request can be satisfied
|
|
|
|
"-" -- means that request can't be satisfied and should wait
|
|
|
|
"0" -- means impossible situation which will trigger assert
|
2009-12-01 14:55:45 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
@note In cases then current context already has "stronger" type
|
|
|
|
of lock on the object it will be automatically granted
|
|
|
|
thanks to usage of the MDL_context::find_ticket() method.
|
|
|
|
*/
|
|
|
|
|
|
|
|
const MDL_lock::bitmap_t
|
|
|
|
MDL_object_lock::m_granted_incompatible[MDL_TYPE_END] =
|
|
|
|
{
|
|
|
|
0,
|
|
|
|
MDL_BIT(MDL_EXCLUSIVE),
|
|
|
|
MDL_BIT(MDL_EXCLUSIVE),
|
|
|
|
MDL_BIT(MDL_EXCLUSIVE) | MDL_BIT(MDL_SHARED_NO_READ_WRITE),
|
|
|
|
MDL_BIT(MDL_EXCLUSIVE) | MDL_BIT(MDL_SHARED_NO_READ_WRITE) |
|
|
|
|
MDL_BIT(MDL_SHARED_NO_WRITE),
|
|
|
|
MDL_BIT(MDL_EXCLUSIVE) | MDL_BIT(MDL_SHARED_NO_READ_WRITE) |
|
|
|
|
MDL_BIT(MDL_SHARED_NO_WRITE) | MDL_BIT(MDL_SHARED_WRITE),
|
|
|
|
MDL_BIT(MDL_EXCLUSIVE) | MDL_BIT(MDL_SHARED_NO_READ_WRITE) |
|
|
|
|
MDL_BIT(MDL_SHARED_NO_WRITE) | MDL_BIT(MDL_SHARED_WRITE) |
|
|
|
|
MDL_BIT(MDL_SHARED_READ),
|
|
|
|
MDL_BIT(MDL_EXCLUSIVE) | MDL_BIT(MDL_SHARED_NO_READ_WRITE) |
|
|
|
|
MDL_BIT(MDL_SHARED_NO_WRITE) | MDL_BIT(MDL_SHARED_WRITE) |
|
|
|
|
MDL_BIT(MDL_SHARED_READ) | MDL_BIT(MDL_SHARED_HIGH_PRIO) |
|
|
|
|
MDL_BIT(MDL_SHARED)
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
|
|
const MDL_lock::bitmap_t
|
|
|
|
MDL_object_lock::m_waiting_incompatible[MDL_TYPE_END] =
|
|
|
|
{
|
|
|
|
0,
|
|
|
|
MDL_BIT(MDL_EXCLUSIVE),
|
|
|
|
0,
|
|
|
|
MDL_BIT(MDL_EXCLUSIVE) | MDL_BIT(MDL_SHARED_NO_READ_WRITE),
|
|
|
|
MDL_BIT(MDL_EXCLUSIVE) | MDL_BIT(MDL_SHARED_NO_READ_WRITE) |
|
|
|
|
MDL_BIT(MDL_SHARED_NO_WRITE),
|
|
|
|
MDL_BIT(MDL_EXCLUSIVE),
|
|
|
|
MDL_BIT(MDL_EXCLUSIVE),
|
|
|
|
0
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
Check if request for the metadata lock can be satisfied given its
|
|
|
|
current state.
|
|
|
|
|
|
|
|
@param type_arg The requested lock type.
|
|
|
|
@param requestor_ctx The MDL context of the requestor.
|
|
|
|
|
|
|
|
@retval TRUE Lock request can be satisfied
|
|
|
|
@retval FALSE There is some conflicting lock.
|
|
|
|
|
|
|
|
@note In cases then current context already has "stronger" type
|
|
|
|
of lock on the object it will be automatically granted
|
|
|
|
thanks to usage of the MDL_context::find_ticket() method.
|
2009-12-01 14:55:45 +01:00
|
|
|
*/
|
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
bool
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
MDL_lock::can_grant_lock(enum_mdl_type type_arg,
|
|
|
|
MDL_context *requestor_ctx) const
|
2009-12-01 14:55:45 +01:00
|
|
|
{
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
bool can_grant= FALSE;
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
bitmap_t waiting_incompat_map= incompatible_waiting_types_bitmap()[type_arg];
|
|
|
|
bitmap_t granted_incompat_map= incompatible_granted_types_bitmap()[type_arg];
|
|
|
|
/*
|
|
|
|
New lock request can be satisfied iff:
|
|
|
|
- There are no incompatible types of satisfied requests
|
|
|
|
in other contexts
|
|
|
|
- There are no waiting requests which have higher priority
|
|
|
|
than this request.
|
|
|
|
*/
|
|
|
|
if (! (m_waiting.bitmap() & waiting_incompat_map))
|
|
|
|
{
|
|
|
|
if (! (m_granted.bitmap() & granted_incompat_map))
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
can_grant= TRUE;
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
else
|
2009-12-01 14:55:45 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
Ticket_iterator it(m_granted);
|
|
|
|
MDL_ticket *ticket;
|
2009-12-01 14:55:45 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
/* Check that the incompatible lock belongs to some other context. */
|
|
|
|
while ((ticket= it++))
|
2009-12-01 14:55:45 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
if (ticket->get_ctx() != requestor_ctx &&
|
|
|
|
ticket->is_incompatible_when_granted(type_arg))
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
break;
|
2009-12-01 14:55:45 +01:00
|
|
|
}
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
if (ticket == NULL) /* Incompatible locks are our own. */
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
can_grant= TRUE;
|
2009-12-01 14:55:45 +01:00
|
|
|
}
|
|
|
|
}
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
return can_grant;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
/** Remove a ticket from waiting or pending queue and wakeup up waiters. */
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
void MDL_lock::remove_ticket(Ticket_list MDL_lock::*list, MDL_ticket *ticket)
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
rw_wrlock(&m_rwlock);
|
|
|
|
(this->*list).remove_ticket(ticket);
|
|
|
|
if (is_empty())
|
|
|
|
mdl_locks.remove(this);
|
|
|
|
else
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
/*
|
|
|
|
There can be some contexts waiting to acquire a lock
|
|
|
|
which now might be able to do it. Wake them up!
|
|
|
|
*/
|
|
|
|
wake_up_waiters();
|
|
|
|
rw_unlock(&m_rwlock);
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
}
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
Check if we have any pending locks which conflict with existing
|
|
|
|
shared lock.
|
|
|
|
|
|
|
|
@pre The ticket must match an acquired lock.
|
|
|
|
|
|
|
|
@return TRUE if there is a conflicting lock request, FALSE otherwise.
|
|
|
|
*/
|
|
|
|
|
|
|
|
bool MDL_lock::has_pending_conflicting_lock(enum_mdl_type type)
|
|
|
|
{
|
|
|
|
bool result;
|
|
|
|
|
2010-02-02 00:22:16 +01:00
|
|
|
mysql_mutex_assert_not_owner(&LOCK_open);
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
|
|
|
|
rw_rdlock(&m_rwlock);
|
|
|
|
result= (m_waiting.bitmap() & incompatible_granted_types_bitmap()[type]);
|
|
|
|
rw_unlock(&m_rwlock);
|
|
|
|
return result;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
Check if ticket represents metadata lock of "stronger" or equal type
|
|
|
|
than specified one. I.e. if metadata lock represented by ticket won't
|
|
|
|
allow any of locks which are not allowed by specified type of lock.
|
|
|
|
|
|
|
|
@return TRUE if ticket has stronger or equal type
|
|
|
|
FALSE otherwise.
|
|
|
|
*/
|
|
|
|
|
|
|
|
bool MDL_ticket::has_stronger_or_equal_type(enum_mdl_type type) const
|
|
|
|
{
|
|
|
|
const MDL_lock::bitmap_t *
|
|
|
|
granted_incompat_map= m_lock->incompatible_granted_types_bitmap();
|
|
|
|
|
|
|
|
return ! (granted_incompat_map[type] & ~(granted_incompat_map[m_type]));
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
bool MDL_ticket::is_incompatible_when_granted(enum_mdl_type type) const
|
|
|
|
{
|
|
|
|
return (MDL_BIT(m_type) &
|
|
|
|
m_lock->incompatible_granted_types_bitmap()[type]);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
bool MDL_ticket::is_incompatible_when_waiting(enum_mdl_type type) const
|
|
|
|
{
|
|
|
|
return (MDL_BIT(m_type) &
|
|
|
|
m_lock->incompatible_waiting_types_bitmap()[type]);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
/**
|
2009-12-04 00:34:19 +01:00
|
|
|
Check whether the context already holds a compatible lock ticket
|
Backport of:
----------------------------------------------------------
revno: 2617.69.20
committer: Konstantin Osipov <kostja@sun.com>
branch nick: 5.4-4284-1-assert
timestamp: Thu 2009-08-13 18:29:55 +0400
message:
WL#4284 "Transactional DDL locking"
A review fix.
Since WL#4284 implementation separated MDL_request and MDL_ticket,
MDL_request becamse a utility object necessary only to get a ticket.
Store it by-value in TABLE_LIST with the intent to merge
MDL_request::key with table_list->table_name and table_list->db
in future.
Change the MDL subsystem to not require MDL_requests to
stay around till close_thread_tables().
Remove the list of requests from the MDL context.
Requests for shared metadata locks acquired in open_tables()
are only used as a list in recover_from_failed_open_table_attempt(),
which calls mdl_context.wait_for_locks() for this list.
To keep such list for recover_from_failed_open_table_attempt(),
introduce a context class (Open_table_context), that collects
all requests.
A lot of minor cleanups and simplications that became possible
with this change.
sql/event_db_repository.cc:
Remove alloc_mdl_requests(). Now MDL_request instance is a member
of TABLE_LIST, and init_one_table() initializes it.
sql/ha_ndbcluster_binlog.cc:
Remove now unnecessary declaration and initialization
of binlog_mdl_request.
sql/lock.cc:
No need to allocate MDL requests in lock_table_names() now.
sql/log.cc:
Use init_one_table() method, remove alloc_mdl_requests(),
which is now unnecessary.
sql/log_event.cc:
No need to allocate mdl_request separately now.
Use init_one_table() method.
sql/log_event_old.cc:
Update to the new signature of close_tables_for_reopen().
sql/mdl.cc:
Update try_acquire_exclusive_lock() to be more easy to use.
Function lock_table_name_if_not_cached() has been removed.
Make acquire_shared_lock() signature consistent with
try_acquire_exclusive_lock() signature.
Remove methods that are no longer used.
Update comments.
sql/mdl.h:
Implement an assignment operator that doesn't
copy MDL_key (MDL_key::operator= is private and
should remain private).
This is a hack to work-around assignment of TABLE_LIST
by value in several places. Such assignments violate
encapsulation, since only perform a shallow copy.
In most cases these assignments are a hack on their own.
sql/mysql_priv.h:
Update signatures of close_thread_tables() and close_tables_for_reopen().
sql/sp.cc:
Allocate TABLE_LIST in thd->mem_root.
Use init_one_table().
sql/sp_head.cc:
Use init_one_table().
Remove thd->locked_tables_root, it's no longer needed.
sql/sql_acl.cc:
Use init_mdl_requests() and init_one_table().
sql/sql_base.cc:
Update to new signatures of try_acquire_shared_lock() and
try_acquire_exclusive_lock().
Remove lock_table_name_if_not_cached().
Fix a bug in open_ltable() that would not return ER_LOCK_DEADLOCK
in case of a failed lock_tables() and a multi-statement
transaction.
Fix a bug in open_and_lock_tables_derived() that would
not return ER_LOCK_DEADLOCK in case of a multi-statement
transaction and a failure of lock_tables().
Move assignment of enum_open_table_action to a method of Open_table_context, a new class that maintains information
for backoff actions.
Minor rearrangements of the code.
Remove alloc_mdl_requests() in functions that work with system
tables: instead the patch ensures that callers always initialize
TABLE_LIST argument.
sql/sql_class.cc:
THD::locked_tables_root is no more.
sql/sql_class.h:
THD::locked_tables_root is no more.
Add a declaration for Open_table_context class.
sql/sql_delete.cc:
Update to use the simplified MDL API.
sql/sql_handler.cc:
TABLE_LIST::mdl_request is stored by-value now.
Ensure that mdl_request.ticket is NULL for every request
that is passed into MDL, to satisfy MDL asserts.
@ sql/sql_help.cc
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_help.cc:
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_insert.cc:
Remove assignment by-value of TABLE_LIST in
TABLEOP_HOOKS. We can't carry over a granted
MDL ticket from one table list to another.
sql/sql_parse.cc:
Change alloc_mdl_requests() -> init_mdl_requests().
@todo We can remove init_mdl_requests() altogether
in some places: all places that call add_table_to_list()
already have mdl requests initialized.
sql/sql_plugin.cc:
Use init_one_table().
THD::locked_tables_root is no more.
sql/sql_servers.cc:
Use init_one_table().
sql/sql_show.cc:
Update acquire_high_priority_shared_lock() to use
TABLE_LIST::mdl_request rather than allocate an own.
Fix get_trigger_table_impl() to use init_one_table(),
check for out of memory, follow the coding style.
sql/sql_table.cc:
Update to work with TABLE_LIST::mdl_request by-value.
Remove lock_table_name_if_not_cached().
The code that used to delegate to it is quite simple and
concise without it now.
sql/sql_udf.cc:
Use init_one_table().
sql/sql_update.cc:
Update to use the new signature of close_tables_for_reopen().
sql/table.cc:
Move re-setting of mdl_requests for prepared statements
and stored procedures from close_thread_tables() to
reinit_stmt_before_use().
Change alloc_mdl_requests() to init_mdl_requests().
init_mdl_requests() is a hack that can't be deleted
until we don't have a list-aware TABLE_LIST constructor.
Hopefully its use will be minimal
sql/table.h:
Change alloc_mdl_requests() to init_mdl_requests()
TABLE_LIST::mdl_request is stored by value.
sql/tztime.cc:
We no longer initialize mdl requests in open_system_tables_for*()
functions. Move this initialization closer to initialization
of the rest of TABLE_LIST members.
storage/myisammrg/ha_myisammrg.cc:
Simplify mdl_request initialization.
2009-12-08 10:57:07 +01:00
|
|
|
on an object.
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
Start searching the transactional locks. If not
|
|
|
|
found in the list of transactional locks, look at LOCK TABLES
|
|
|
|
and HANDLER locks.
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
@param mdl_request Lock request object for lock to be acquired
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
@param[out] is_transactional FALSE if we pass beyond m_trans_sentinel
|
|
|
|
while searching for ticket, otherwise TRUE.
|
|
|
|
|
|
|
|
@note Tickets which correspond to lock types "stronger" than one
|
|
|
|
being requested are also considered compatible.
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
@return A pointer to the lock ticket for the object or NULL otherwise.
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
*/
|
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
MDL_ticket *
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
MDL_context::find_ticket(MDL_request *mdl_request,
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
bool *is_transactional)
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
{
|
2009-12-04 00:52:05 +01:00
|
|
|
MDL_ticket *ticket;
|
|
|
|
Ticket_iterator it(m_tickets);
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
*is_transactional= TRUE;
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
while ((ticket= it++))
|
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
if (ticket == m_trans_sentinel)
|
|
|
|
*is_transactional= FALSE;
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
if (mdl_request->key.is_equal(&ticket->m_lock->key) &&
|
|
|
|
ticket->has_stronger_or_equal_type(mdl_request->type))
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
return ticket;
|
2009-12-01 14:55:45 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
/**
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
Acquire one lock with waiting for conflicting locks to go away if needed.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
@note This is an internal method which should not be used outside of MDL
|
|
|
|
subsystem as in most cases simply waiting for conflicting locks to
|
|
|
|
go away will lead to deadlock.
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
|
|
|
@param mdl_request [in/out] Lock request object for lock to be acquired
|
|
|
|
|
|
|
|
@retval FALSE Success. MDL_request::ticket points to the ticket
|
|
|
|
for the lock.
|
|
|
|
@retval TRUE Failure (Out of resources or waiting is aborted),
|
|
|
|
*/
|
|
|
|
|
|
|
|
bool
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
MDL_context::acquire_lock(MDL_request *mdl_request)
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
return acquire_lock_impl(mdl_request);
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
Try to acquire one lock.
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
Unlike exclusive locks, shared locks are acquired one by
|
|
|
|
one. This is interface is chosen to simplify introduction of
|
|
|
|
the new locking API to the system. MDL_context::try_acquire_lock()
|
|
|
|
is currently used from open_table(), and there we have only one
|
|
|
|
table to work with.
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
This function may also be used to try to acquire an exclusive
|
|
|
|
lock on a destination table, by ALTER TABLE ... RENAME.
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
Returns immediately without any side effect if encounters a lock
|
|
|
|
conflict. Otherwise takes the lock.
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
FIXME: Compared to lock_table_name_if_not_cached() (from 5.1)
|
|
|
|
it gives slightly more false negatives.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Backport of:
----------------------------------------------------------
revno: 2617.69.20
committer: Konstantin Osipov <kostja@sun.com>
branch nick: 5.4-4284-1-assert
timestamp: Thu 2009-08-13 18:29:55 +0400
message:
WL#4284 "Transactional DDL locking"
A review fix.
Since WL#4284 implementation separated MDL_request and MDL_ticket,
MDL_request becamse a utility object necessary only to get a ticket.
Store it by-value in TABLE_LIST with the intent to merge
MDL_request::key with table_list->table_name and table_list->db
in future.
Change the MDL subsystem to not require MDL_requests to
stay around till close_thread_tables().
Remove the list of requests from the MDL context.
Requests for shared metadata locks acquired in open_tables()
are only used as a list in recover_from_failed_open_table_attempt(),
which calls mdl_context.wait_for_locks() for this list.
To keep such list for recover_from_failed_open_table_attempt(),
introduce a context class (Open_table_context), that collects
all requests.
A lot of minor cleanups and simplications that became possible
with this change.
sql/event_db_repository.cc:
Remove alloc_mdl_requests(). Now MDL_request instance is a member
of TABLE_LIST, and init_one_table() initializes it.
sql/ha_ndbcluster_binlog.cc:
Remove now unnecessary declaration and initialization
of binlog_mdl_request.
sql/lock.cc:
No need to allocate MDL requests in lock_table_names() now.
sql/log.cc:
Use init_one_table() method, remove alloc_mdl_requests(),
which is now unnecessary.
sql/log_event.cc:
No need to allocate mdl_request separately now.
Use init_one_table() method.
sql/log_event_old.cc:
Update to the new signature of close_tables_for_reopen().
sql/mdl.cc:
Update try_acquire_exclusive_lock() to be more easy to use.
Function lock_table_name_if_not_cached() has been removed.
Make acquire_shared_lock() signature consistent with
try_acquire_exclusive_lock() signature.
Remove methods that are no longer used.
Update comments.
sql/mdl.h:
Implement an assignment operator that doesn't
copy MDL_key (MDL_key::operator= is private and
should remain private).
This is a hack to work-around assignment of TABLE_LIST
by value in several places. Such assignments violate
encapsulation, since only perform a shallow copy.
In most cases these assignments are a hack on their own.
sql/mysql_priv.h:
Update signatures of close_thread_tables() and close_tables_for_reopen().
sql/sp.cc:
Allocate TABLE_LIST in thd->mem_root.
Use init_one_table().
sql/sp_head.cc:
Use init_one_table().
Remove thd->locked_tables_root, it's no longer needed.
sql/sql_acl.cc:
Use init_mdl_requests() and init_one_table().
sql/sql_base.cc:
Update to new signatures of try_acquire_shared_lock() and
try_acquire_exclusive_lock().
Remove lock_table_name_if_not_cached().
Fix a bug in open_ltable() that would not return ER_LOCK_DEADLOCK
in case of a failed lock_tables() and a multi-statement
transaction.
Fix a bug in open_and_lock_tables_derived() that would
not return ER_LOCK_DEADLOCK in case of a multi-statement
transaction and a failure of lock_tables().
Move assignment of enum_open_table_action to a method of Open_table_context, a new class that maintains information
for backoff actions.
Minor rearrangements of the code.
Remove alloc_mdl_requests() in functions that work with system
tables: instead the patch ensures that callers always initialize
TABLE_LIST argument.
sql/sql_class.cc:
THD::locked_tables_root is no more.
sql/sql_class.h:
THD::locked_tables_root is no more.
Add a declaration for Open_table_context class.
sql/sql_delete.cc:
Update to use the simplified MDL API.
sql/sql_handler.cc:
TABLE_LIST::mdl_request is stored by-value now.
Ensure that mdl_request.ticket is NULL for every request
that is passed into MDL, to satisfy MDL asserts.
@ sql/sql_help.cc
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_help.cc:
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_insert.cc:
Remove assignment by-value of TABLE_LIST in
TABLEOP_HOOKS. We can't carry over a granted
MDL ticket from one table list to another.
sql/sql_parse.cc:
Change alloc_mdl_requests() -> init_mdl_requests().
@todo We can remove init_mdl_requests() altogether
in some places: all places that call add_table_to_list()
already have mdl requests initialized.
sql/sql_plugin.cc:
Use init_one_table().
THD::locked_tables_root is no more.
sql/sql_servers.cc:
Use init_one_table().
sql/sql_show.cc:
Update acquire_high_priority_shared_lock() to use
TABLE_LIST::mdl_request rather than allocate an own.
Fix get_trigger_table_impl() to use init_one_table(),
check for out of memory, follow the coding style.
sql/sql_table.cc:
Update to work with TABLE_LIST::mdl_request by-value.
Remove lock_table_name_if_not_cached().
The code that used to delegate to it is quite simple and
concise without it now.
sql/sql_udf.cc:
Use init_one_table().
sql/sql_update.cc:
Update to use the new signature of close_tables_for_reopen().
sql/table.cc:
Move re-setting of mdl_requests for prepared statements
and stored procedures from close_thread_tables() to
reinit_stmt_before_use().
Change alloc_mdl_requests() to init_mdl_requests().
init_mdl_requests() is a hack that can't be deleted
until we don't have a list-aware TABLE_LIST constructor.
Hopefully its use will be minimal
sql/table.h:
Change alloc_mdl_requests() to init_mdl_requests()
TABLE_LIST::mdl_request is stored by value.
sql/tztime.cc:
We no longer initialize mdl requests in open_system_tables_for*()
functions. Move this initialization closer to initialization
of the rest of TABLE_LIST members.
storage/myisammrg/ha_myisammrg.cc:
Simplify mdl_request initialization.
2009-12-08 10:57:07 +01:00
|
|
|
@param mdl_request [in/out] Lock request object for lock to be acquired
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Backport of:
----------------------------------------------------------
revno: 2617.69.20
committer: Konstantin Osipov <kostja@sun.com>
branch nick: 5.4-4284-1-assert
timestamp: Thu 2009-08-13 18:29:55 +0400
message:
WL#4284 "Transactional DDL locking"
A review fix.
Since WL#4284 implementation separated MDL_request and MDL_ticket,
MDL_request becamse a utility object necessary only to get a ticket.
Store it by-value in TABLE_LIST with the intent to merge
MDL_request::key with table_list->table_name and table_list->db
in future.
Change the MDL subsystem to not require MDL_requests to
stay around till close_thread_tables().
Remove the list of requests from the MDL context.
Requests for shared metadata locks acquired in open_tables()
are only used as a list in recover_from_failed_open_table_attempt(),
which calls mdl_context.wait_for_locks() for this list.
To keep such list for recover_from_failed_open_table_attempt(),
introduce a context class (Open_table_context), that collects
all requests.
A lot of minor cleanups and simplications that became possible
with this change.
sql/event_db_repository.cc:
Remove alloc_mdl_requests(). Now MDL_request instance is a member
of TABLE_LIST, and init_one_table() initializes it.
sql/ha_ndbcluster_binlog.cc:
Remove now unnecessary declaration and initialization
of binlog_mdl_request.
sql/lock.cc:
No need to allocate MDL requests in lock_table_names() now.
sql/log.cc:
Use init_one_table() method, remove alloc_mdl_requests(),
which is now unnecessary.
sql/log_event.cc:
No need to allocate mdl_request separately now.
Use init_one_table() method.
sql/log_event_old.cc:
Update to the new signature of close_tables_for_reopen().
sql/mdl.cc:
Update try_acquire_exclusive_lock() to be more easy to use.
Function lock_table_name_if_not_cached() has been removed.
Make acquire_shared_lock() signature consistent with
try_acquire_exclusive_lock() signature.
Remove methods that are no longer used.
Update comments.
sql/mdl.h:
Implement an assignment operator that doesn't
copy MDL_key (MDL_key::operator= is private and
should remain private).
This is a hack to work-around assignment of TABLE_LIST
by value in several places. Such assignments violate
encapsulation, since only perform a shallow copy.
In most cases these assignments are a hack on their own.
sql/mysql_priv.h:
Update signatures of close_thread_tables() and close_tables_for_reopen().
sql/sp.cc:
Allocate TABLE_LIST in thd->mem_root.
Use init_one_table().
sql/sp_head.cc:
Use init_one_table().
Remove thd->locked_tables_root, it's no longer needed.
sql/sql_acl.cc:
Use init_mdl_requests() and init_one_table().
sql/sql_base.cc:
Update to new signatures of try_acquire_shared_lock() and
try_acquire_exclusive_lock().
Remove lock_table_name_if_not_cached().
Fix a bug in open_ltable() that would not return ER_LOCK_DEADLOCK
in case of a failed lock_tables() and a multi-statement
transaction.
Fix a bug in open_and_lock_tables_derived() that would
not return ER_LOCK_DEADLOCK in case of a multi-statement
transaction and a failure of lock_tables().
Move assignment of enum_open_table_action to a method of Open_table_context, a new class that maintains information
for backoff actions.
Minor rearrangements of the code.
Remove alloc_mdl_requests() in functions that work with system
tables: instead the patch ensures that callers always initialize
TABLE_LIST argument.
sql/sql_class.cc:
THD::locked_tables_root is no more.
sql/sql_class.h:
THD::locked_tables_root is no more.
Add a declaration for Open_table_context class.
sql/sql_delete.cc:
Update to use the simplified MDL API.
sql/sql_handler.cc:
TABLE_LIST::mdl_request is stored by-value now.
Ensure that mdl_request.ticket is NULL for every request
that is passed into MDL, to satisfy MDL asserts.
@ sql/sql_help.cc
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_help.cc:
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_insert.cc:
Remove assignment by-value of TABLE_LIST in
TABLEOP_HOOKS. We can't carry over a granted
MDL ticket from one table list to another.
sql/sql_parse.cc:
Change alloc_mdl_requests() -> init_mdl_requests().
@todo We can remove init_mdl_requests() altogether
in some places: all places that call add_table_to_list()
already have mdl requests initialized.
sql/sql_plugin.cc:
Use init_one_table().
THD::locked_tables_root is no more.
sql/sql_servers.cc:
Use init_one_table().
sql/sql_show.cc:
Update acquire_high_priority_shared_lock() to use
TABLE_LIST::mdl_request rather than allocate an own.
Fix get_trigger_table_impl() to use init_one_table(),
check for out of memory, follow the coding style.
sql/sql_table.cc:
Update to work with TABLE_LIST::mdl_request by-value.
Remove lock_table_name_if_not_cached().
The code that used to delegate to it is quite simple and
concise without it now.
sql/sql_udf.cc:
Use init_one_table().
sql/sql_update.cc:
Update to use the new signature of close_tables_for_reopen().
sql/table.cc:
Move re-setting of mdl_requests for prepared statements
and stored procedures from close_thread_tables() to
reinit_stmt_before_use().
Change alloc_mdl_requests() to init_mdl_requests().
init_mdl_requests() is a hack that can't be deleted
until we don't have a list-aware TABLE_LIST constructor.
Hopefully its use will be minimal
sql/table.h:
Change alloc_mdl_requests() to init_mdl_requests()
TABLE_LIST::mdl_request is stored by value.
sql/tztime.cc:
We no longer initialize mdl requests in open_system_tables_for*()
functions. Move this initialization closer to initialization
of the rest of TABLE_LIST members.
storage/myisammrg/ha_myisammrg.cc:
Simplify mdl_request initialization.
2009-12-08 10:57:07 +01:00
|
|
|
@retval FALSE Success. The lock may have not been acquired.
|
|
|
|
Check the ticket, if it's NULL, a conflicting lock
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
exists and another attempt should be made after releasing
|
|
|
|
all current locks and waiting for conflicting lock go
|
|
|
|
away (using MDL_context::wait_for_lock()).
|
Backport of:
----------------------------------------------------------
revno: 2617.69.20
committer: Konstantin Osipov <kostja@sun.com>
branch nick: 5.4-4284-1-assert
timestamp: Thu 2009-08-13 18:29:55 +0400
message:
WL#4284 "Transactional DDL locking"
A review fix.
Since WL#4284 implementation separated MDL_request and MDL_ticket,
MDL_request becamse a utility object necessary only to get a ticket.
Store it by-value in TABLE_LIST with the intent to merge
MDL_request::key with table_list->table_name and table_list->db
in future.
Change the MDL subsystem to not require MDL_requests to
stay around till close_thread_tables().
Remove the list of requests from the MDL context.
Requests for shared metadata locks acquired in open_tables()
are only used as a list in recover_from_failed_open_table_attempt(),
which calls mdl_context.wait_for_locks() for this list.
To keep such list for recover_from_failed_open_table_attempt(),
introduce a context class (Open_table_context), that collects
all requests.
A lot of minor cleanups and simplications that became possible
with this change.
sql/event_db_repository.cc:
Remove alloc_mdl_requests(). Now MDL_request instance is a member
of TABLE_LIST, and init_one_table() initializes it.
sql/ha_ndbcluster_binlog.cc:
Remove now unnecessary declaration and initialization
of binlog_mdl_request.
sql/lock.cc:
No need to allocate MDL requests in lock_table_names() now.
sql/log.cc:
Use init_one_table() method, remove alloc_mdl_requests(),
which is now unnecessary.
sql/log_event.cc:
No need to allocate mdl_request separately now.
Use init_one_table() method.
sql/log_event_old.cc:
Update to the new signature of close_tables_for_reopen().
sql/mdl.cc:
Update try_acquire_exclusive_lock() to be more easy to use.
Function lock_table_name_if_not_cached() has been removed.
Make acquire_shared_lock() signature consistent with
try_acquire_exclusive_lock() signature.
Remove methods that are no longer used.
Update comments.
sql/mdl.h:
Implement an assignment operator that doesn't
copy MDL_key (MDL_key::operator= is private and
should remain private).
This is a hack to work-around assignment of TABLE_LIST
by value in several places. Such assignments violate
encapsulation, since only perform a shallow copy.
In most cases these assignments are a hack on their own.
sql/mysql_priv.h:
Update signatures of close_thread_tables() and close_tables_for_reopen().
sql/sp.cc:
Allocate TABLE_LIST in thd->mem_root.
Use init_one_table().
sql/sp_head.cc:
Use init_one_table().
Remove thd->locked_tables_root, it's no longer needed.
sql/sql_acl.cc:
Use init_mdl_requests() and init_one_table().
sql/sql_base.cc:
Update to new signatures of try_acquire_shared_lock() and
try_acquire_exclusive_lock().
Remove lock_table_name_if_not_cached().
Fix a bug in open_ltable() that would not return ER_LOCK_DEADLOCK
in case of a failed lock_tables() and a multi-statement
transaction.
Fix a bug in open_and_lock_tables_derived() that would
not return ER_LOCK_DEADLOCK in case of a multi-statement
transaction and a failure of lock_tables().
Move assignment of enum_open_table_action to a method of Open_table_context, a new class that maintains information
for backoff actions.
Minor rearrangements of the code.
Remove alloc_mdl_requests() in functions that work with system
tables: instead the patch ensures that callers always initialize
TABLE_LIST argument.
sql/sql_class.cc:
THD::locked_tables_root is no more.
sql/sql_class.h:
THD::locked_tables_root is no more.
Add a declaration for Open_table_context class.
sql/sql_delete.cc:
Update to use the simplified MDL API.
sql/sql_handler.cc:
TABLE_LIST::mdl_request is stored by-value now.
Ensure that mdl_request.ticket is NULL for every request
that is passed into MDL, to satisfy MDL asserts.
@ sql/sql_help.cc
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_help.cc:
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_insert.cc:
Remove assignment by-value of TABLE_LIST in
TABLEOP_HOOKS. We can't carry over a granted
MDL ticket from one table list to another.
sql/sql_parse.cc:
Change alloc_mdl_requests() -> init_mdl_requests().
@todo We can remove init_mdl_requests() altogether
in some places: all places that call add_table_to_list()
already have mdl requests initialized.
sql/sql_plugin.cc:
Use init_one_table().
THD::locked_tables_root is no more.
sql/sql_servers.cc:
Use init_one_table().
sql/sql_show.cc:
Update acquire_high_priority_shared_lock() to use
TABLE_LIST::mdl_request rather than allocate an own.
Fix get_trigger_table_impl() to use init_one_table(),
check for out of memory, follow the coding style.
sql/sql_table.cc:
Update to work with TABLE_LIST::mdl_request by-value.
Remove lock_table_name_if_not_cached().
The code that used to delegate to it is quite simple and
concise without it now.
sql/sql_udf.cc:
Use init_one_table().
sql/sql_update.cc:
Update to use the new signature of close_tables_for_reopen().
sql/table.cc:
Move re-setting of mdl_requests for prepared statements
and stored procedures from close_thread_tables() to
reinit_stmt_before_use().
Change alloc_mdl_requests() to init_mdl_requests().
init_mdl_requests() is a hack that can't be deleted
until we don't have a list-aware TABLE_LIST constructor.
Hopefully its use will be minimal
sql/table.h:
Change alloc_mdl_requests() to init_mdl_requests()
TABLE_LIST::mdl_request is stored by value.
sql/tztime.cc:
We no longer initialize mdl requests in open_system_tables_for*()
functions. Move this initialization closer to initialization
of the rest of TABLE_LIST members.
storage/myisammrg/ha_myisammrg.cc:
Simplify mdl_request initialization.
2009-12-08 10:57:07 +01:00
|
|
|
@retval TRUE Out of resources, an error has been reported.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
*/
|
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
bool
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
MDL_context::try_acquire_lock(MDL_request *mdl_request)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
2009-12-04 00:52:05 +01:00
|
|
|
MDL_lock *lock;
|
|
|
|
MDL_key *key= &mdl_request->key;
|
|
|
|
MDL_ticket *ticket;
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
bool is_transactional;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
DBUG_ASSERT(mdl_request->type < MDL_SHARED_NO_WRITE ||
|
|
|
|
(is_lock_owner(MDL_key::GLOBAL, "", "",
|
|
|
|
MDL_INTENTION_EXCLUSIVE)));
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
DBUG_ASSERT(mdl_request->ticket == NULL);
|
2009-12-01 16:20:43 +01:00
|
|
|
|
Backport of:
----------------------------------------------------------
revno: 2617.69.20
committer: Konstantin Osipov <kostja@sun.com>
branch nick: 5.4-4284-1-assert
timestamp: Thu 2009-08-13 18:29:55 +0400
message:
WL#4284 "Transactional DDL locking"
A review fix.
Since WL#4284 implementation separated MDL_request and MDL_ticket,
MDL_request becamse a utility object necessary only to get a ticket.
Store it by-value in TABLE_LIST with the intent to merge
MDL_request::key with table_list->table_name and table_list->db
in future.
Change the MDL subsystem to not require MDL_requests to
stay around till close_thread_tables().
Remove the list of requests from the MDL context.
Requests for shared metadata locks acquired in open_tables()
are only used as a list in recover_from_failed_open_table_attempt(),
which calls mdl_context.wait_for_locks() for this list.
To keep such list for recover_from_failed_open_table_attempt(),
introduce a context class (Open_table_context), that collects
all requests.
A lot of minor cleanups and simplications that became possible
with this change.
sql/event_db_repository.cc:
Remove alloc_mdl_requests(). Now MDL_request instance is a member
of TABLE_LIST, and init_one_table() initializes it.
sql/ha_ndbcluster_binlog.cc:
Remove now unnecessary declaration and initialization
of binlog_mdl_request.
sql/lock.cc:
No need to allocate MDL requests in lock_table_names() now.
sql/log.cc:
Use init_one_table() method, remove alloc_mdl_requests(),
which is now unnecessary.
sql/log_event.cc:
No need to allocate mdl_request separately now.
Use init_one_table() method.
sql/log_event_old.cc:
Update to the new signature of close_tables_for_reopen().
sql/mdl.cc:
Update try_acquire_exclusive_lock() to be more easy to use.
Function lock_table_name_if_not_cached() has been removed.
Make acquire_shared_lock() signature consistent with
try_acquire_exclusive_lock() signature.
Remove methods that are no longer used.
Update comments.
sql/mdl.h:
Implement an assignment operator that doesn't
copy MDL_key (MDL_key::operator= is private and
should remain private).
This is a hack to work-around assignment of TABLE_LIST
by value in several places. Such assignments violate
encapsulation, since only perform a shallow copy.
In most cases these assignments are a hack on their own.
sql/mysql_priv.h:
Update signatures of close_thread_tables() and close_tables_for_reopen().
sql/sp.cc:
Allocate TABLE_LIST in thd->mem_root.
Use init_one_table().
sql/sp_head.cc:
Use init_one_table().
Remove thd->locked_tables_root, it's no longer needed.
sql/sql_acl.cc:
Use init_mdl_requests() and init_one_table().
sql/sql_base.cc:
Update to new signatures of try_acquire_shared_lock() and
try_acquire_exclusive_lock().
Remove lock_table_name_if_not_cached().
Fix a bug in open_ltable() that would not return ER_LOCK_DEADLOCK
in case of a failed lock_tables() and a multi-statement
transaction.
Fix a bug in open_and_lock_tables_derived() that would
not return ER_LOCK_DEADLOCK in case of a multi-statement
transaction and a failure of lock_tables().
Move assignment of enum_open_table_action to a method of Open_table_context, a new class that maintains information
for backoff actions.
Minor rearrangements of the code.
Remove alloc_mdl_requests() in functions that work with system
tables: instead the patch ensures that callers always initialize
TABLE_LIST argument.
sql/sql_class.cc:
THD::locked_tables_root is no more.
sql/sql_class.h:
THD::locked_tables_root is no more.
Add a declaration for Open_table_context class.
sql/sql_delete.cc:
Update to use the simplified MDL API.
sql/sql_handler.cc:
TABLE_LIST::mdl_request is stored by-value now.
Ensure that mdl_request.ticket is NULL for every request
that is passed into MDL, to satisfy MDL asserts.
@ sql/sql_help.cc
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_help.cc:
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_insert.cc:
Remove assignment by-value of TABLE_LIST in
TABLEOP_HOOKS. We can't carry over a granted
MDL ticket from one table list to another.
sql/sql_parse.cc:
Change alloc_mdl_requests() -> init_mdl_requests().
@todo We can remove init_mdl_requests() altogether
in some places: all places that call add_table_to_list()
already have mdl requests initialized.
sql/sql_plugin.cc:
Use init_one_table().
THD::locked_tables_root is no more.
sql/sql_servers.cc:
Use init_one_table().
sql/sql_show.cc:
Update acquire_high_priority_shared_lock() to use
TABLE_LIST::mdl_request rather than allocate an own.
Fix get_trigger_table_impl() to use init_one_table(),
check for out of memory, follow the coding style.
sql/sql_table.cc:
Update to work with TABLE_LIST::mdl_request by-value.
Remove lock_table_name_if_not_cached().
The code that used to delegate to it is quite simple and
concise without it now.
sql/sql_udf.cc:
Use init_one_table().
sql/sql_update.cc:
Update to use the new signature of close_tables_for_reopen().
sql/table.cc:
Move re-setting of mdl_requests for prepared statements
and stored procedures from close_thread_tables() to
reinit_stmt_before_use().
Change alloc_mdl_requests() to init_mdl_requests().
init_mdl_requests() is a hack that can't be deleted
until we don't have a list-aware TABLE_LIST constructor.
Hopefully its use will be minimal
sql/table.h:
Change alloc_mdl_requests() to init_mdl_requests()
TABLE_LIST::mdl_request is stored by value.
sql/tztime.cc:
We no longer initialize mdl requests in open_system_tables_for*()
functions. Move this initialization closer to initialization
of the rest of TABLE_LIST members.
storage/myisammrg/ha_myisammrg.cc:
Simplify mdl_request initialization.
2009-12-08 10:57:07 +01:00
|
|
|
/* Don't take chances in production. */
|
|
|
|
mdl_request->ticket= NULL;
|
2010-02-02 00:22:16 +01:00
|
|
|
mysql_mutex_assert_not_owner(&LOCK_open);
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
/*
|
|
|
|
Check whether the context already holds a shared lock on the object,
|
|
|
|
and if so, grant the request.
|
|
|
|
*/
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
if ((ticket= find_ticket(mdl_request, &is_transactional)))
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
DBUG_ASSERT(ticket->m_lock);
|
|
|
|
DBUG_ASSERT(ticket->m_type >= mdl_request->type);
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
/*
|
|
|
|
If the request is for a transactional lock, and we found
|
|
|
|
a transactional lock, just reuse the found ticket.
|
|
|
|
|
|
|
|
It's possible that we found a transactional lock,
|
|
|
|
but the request is for a HANDLER lock. In that case HANDLER
|
|
|
|
code will clone the ticket (see below why it's needed).
|
|
|
|
|
|
|
|
If the request is for a transactional lock, and we found
|
|
|
|
a HANDLER lock, create a copy, to make sure that when user
|
|
|
|
does HANDLER CLOSE, the transactional lock is not released.
|
|
|
|
|
|
|
|
If the request is for a handler lock, and we found a
|
|
|
|
HANDLER lock, also do the clone. HANDLER CLOSE for one alias
|
|
|
|
should not release the lock on the table HANDLER opened through
|
|
|
|
a different alias.
|
|
|
|
*/
|
2009-12-04 00:52:05 +01:00
|
|
|
mdl_request->ticket= ticket;
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
if (!is_transactional && clone_ticket(mdl_request))
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
{
|
|
|
|
/* Clone failed. */
|
|
|
|
mdl_request->ticket= NULL;
|
|
|
|
return TRUE;
|
|
|
|
}
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
if (!(ticket= MDL_ticket::create(this, mdl_request->type)))
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
return TRUE;
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
/* The below call implicitly locks MDL_lock::m_rwlock on success. */
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
if (!(lock= mdl_locks.find_or_insert(key)))
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
{
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
MDL_ticket::destroy(ticket);
|
|
|
|
return TRUE;
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
}
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
if (lock->can_grant_lock(mdl_request->type, this))
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
{
|
2010-02-01 15:38:50 +01:00
|
|
|
ticket->m_lock= lock;
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
lock->m_granted.add_ticket(ticket);
|
|
|
|
rw_unlock(&lock->m_rwlock);
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
|
|
|
m_tickets.push_front(ticket);
|
|
|
|
|
|
|
|
mdl_request->ticket= ticket;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
/* We can't get here if we allocated a new lock. */
|
|
|
|
DBUG_ASSERT(! lock->is_empty());
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
rw_unlock(&lock->m_rwlock);
|
2009-12-04 00:52:05 +01:00
|
|
|
MDL_ticket::destroy(ticket);
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
|
|
|
|
Backport of:
----------------------------------------------------------
revno: 2617.69.20
committer: Konstantin Osipov <kostja@sun.com>
branch nick: 5.4-4284-1-assert
timestamp: Thu 2009-08-13 18:29:55 +0400
message:
WL#4284 "Transactional DDL locking"
A review fix.
Since WL#4284 implementation separated MDL_request and MDL_ticket,
MDL_request becamse a utility object necessary only to get a ticket.
Store it by-value in TABLE_LIST with the intent to merge
MDL_request::key with table_list->table_name and table_list->db
in future.
Change the MDL subsystem to not require MDL_requests to
stay around till close_thread_tables().
Remove the list of requests from the MDL context.
Requests for shared metadata locks acquired in open_tables()
are only used as a list in recover_from_failed_open_table_attempt(),
which calls mdl_context.wait_for_locks() for this list.
To keep such list for recover_from_failed_open_table_attempt(),
introduce a context class (Open_table_context), that collects
all requests.
A lot of minor cleanups and simplications that became possible
with this change.
sql/event_db_repository.cc:
Remove alloc_mdl_requests(). Now MDL_request instance is a member
of TABLE_LIST, and init_one_table() initializes it.
sql/ha_ndbcluster_binlog.cc:
Remove now unnecessary declaration and initialization
of binlog_mdl_request.
sql/lock.cc:
No need to allocate MDL requests in lock_table_names() now.
sql/log.cc:
Use init_one_table() method, remove alloc_mdl_requests(),
which is now unnecessary.
sql/log_event.cc:
No need to allocate mdl_request separately now.
Use init_one_table() method.
sql/log_event_old.cc:
Update to the new signature of close_tables_for_reopen().
sql/mdl.cc:
Update try_acquire_exclusive_lock() to be more easy to use.
Function lock_table_name_if_not_cached() has been removed.
Make acquire_shared_lock() signature consistent with
try_acquire_exclusive_lock() signature.
Remove methods that are no longer used.
Update comments.
sql/mdl.h:
Implement an assignment operator that doesn't
copy MDL_key (MDL_key::operator= is private and
should remain private).
This is a hack to work-around assignment of TABLE_LIST
by value in several places. Such assignments violate
encapsulation, since only perform a shallow copy.
In most cases these assignments are a hack on their own.
sql/mysql_priv.h:
Update signatures of close_thread_tables() and close_tables_for_reopen().
sql/sp.cc:
Allocate TABLE_LIST in thd->mem_root.
Use init_one_table().
sql/sp_head.cc:
Use init_one_table().
Remove thd->locked_tables_root, it's no longer needed.
sql/sql_acl.cc:
Use init_mdl_requests() and init_one_table().
sql/sql_base.cc:
Update to new signatures of try_acquire_shared_lock() and
try_acquire_exclusive_lock().
Remove lock_table_name_if_not_cached().
Fix a bug in open_ltable() that would not return ER_LOCK_DEADLOCK
in case of a failed lock_tables() and a multi-statement
transaction.
Fix a bug in open_and_lock_tables_derived() that would
not return ER_LOCK_DEADLOCK in case of a multi-statement
transaction and a failure of lock_tables().
Move assignment of enum_open_table_action to a method of Open_table_context, a new class that maintains information
for backoff actions.
Minor rearrangements of the code.
Remove alloc_mdl_requests() in functions that work with system
tables: instead the patch ensures that callers always initialize
TABLE_LIST argument.
sql/sql_class.cc:
THD::locked_tables_root is no more.
sql/sql_class.h:
THD::locked_tables_root is no more.
Add a declaration for Open_table_context class.
sql/sql_delete.cc:
Update to use the simplified MDL API.
sql/sql_handler.cc:
TABLE_LIST::mdl_request is stored by-value now.
Ensure that mdl_request.ticket is NULL for every request
that is passed into MDL, to satisfy MDL asserts.
@ sql/sql_help.cc
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_help.cc:
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_insert.cc:
Remove assignment by-value of TABLE_LIST in
TABLEOP_HOOKS. We can't carry over a granted
MDL ticket from one table list to another.
sql/sql_parse.cc:
Change alloc_mdl_requests() -> init_mdl_requests().
@todo We can remove init_mdl_requests() altogether
in some places: all places that call add_table_to_list()
already have mdl requests initialized.
sql/sql_plugin.cc:
Use init_one_table().
THD::locked_tables_root is no more.
sql/sql_servers.cc:
Use init_one_table().
sql/sql_show.cc:
Update acquire_high_priority_shared_lock() to use
TABLE_LIST::mdl_request rather than allocate an own.
Fix get_trigger_table_impl() to use init_one_table(),
check for out of memory, follow the coding style.
sql/sql_table.cc:
Update to work with TABLE_LIST::mdl_request by-value.
Remove lock_table_name_if_not_cached().
The code that used to delegate to it is quite simple and
concise without it now.
sql/sql_udf.cc:
Use init_one_table().
sql/sql_update.cc:
Update to use the new signature of close_tables_for_reopen().
sql/table.cc:
Move re-setting of mdl_requests for prepared statements
and stored procedures from close_thread_tables() to
reinit_stmt_before_use().
Change alloc_mdl_requests() to init_mdl_requests().
init_mdl_requests() is a hack that can't be deleted
until we don't have a list-aware TABLE_LIST constructor.
Hopefully its use will be minimal
sql/table.h:
Change alloc_mdl_requests() to init_mdl_requests()
TABLE_LIST::mdl_request is stored by value.
sql/tztime.cc:
We no longer initialize mdl requests in open_system_tables_for*()
functions. Move this initialization closer to initialization
of the rest of TABLE_LIST members.
storage/myisammrg/ha_myisammrg.cc:
Simplify mdl_request initialization.
2009-12-08 10:57:07 +01:00
|
|
|
return FALSE;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
/**
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
Create a copy of a granted ticket.
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
This is used to make sure that HANDLER ticket
|
|
|
|
is never shared with a ticket that belongs to
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
a transaction, so that when we HANDLER CLOSE,
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
we don't release a transactional ticket, and
|
|
|
|
vice versa -- when we COMMIT, we don't mistakenly
|
|
|
|
release a ticket for an open HANDLER.
|
|
|
|
|
|
|
|
@retval TRUE Out of memory.
|
|
|
|
@retval FALSE Success.
|
|
|
|
*/
|
|
|
|
|
|
|
|
bool
|
|
|
|
MDL_context::clone_ticket(MDL_request *mdl_request)
|
|
|
|
{
|
|
|
|
MDL_ticket *ticket;
|
|
|
|
|
2010-02-02 00:22:16 +01:00
|
|
|
mysql_mutex_assert_not_owner(&LOCK_open);
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
/*
|
|
|
|
By submitting mdl_request->type to MDL_ticket::create()
|
|
|
|
we effectively downgrade the cloned lock to the level of
|
|
|
|
the request.
|
|
|
|
*/
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
if (!(ticket= MDL_ticket::create(this, mdl_request->type)))
|
|
|
|
return TRUE;
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
/* clone() is not supposed to be used to get a stronger lock. */
|
|
|
|
DBUG_ASSERT(ticket->m_type <= mdl_request->ticket->m_type);
|
|
|
|
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
ticket->m_lock= mdl_request->ticket->m_lock;
|
|
|
|
mdl_request->ticket= ticket;
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
rw_wrlock(&ticket->m_lock->m_rwlock);
|
|
|
|
ticket->m_lock->m_granted.add_ticket(ticket);
|
|
|
|
rw_unlock(&ticket->m_lock->m_rwlock);
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
|
|
|
m_tickets.push_front(ticket);
|
|
|
|
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
/**
|
|
|
|
Notify a thread holding a shared metadata lock which
|
|
|
|
conflicts with a pending exclusive lock.
|
|
|
|
|
|
|
|
@param thd Current thread context
|
|
|
|
@param conflicting_ticket Conflicting metadata lock
|
|
|
|
*/
|
|
|
|
|
|
|
|
void notify_shared_lock(THD *thd, MDL_ticket *conflicting_ticket)
|
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
/* Only try to abort locks on which we back off. */
|
|
|
|
if (conflicting_ticket->get_type() < MDL_SHARED_NO_WRITE)
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
MDL_context *conflicting_ctx= conflicting_ticket->get_ctx();
|
|
|
|
THD *conflicting_thd= conflicting_ctx->get_thd();
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
DBUG_ASSERT(thd != conflicting_thd); /* Self-deadlock */
|
|
|
|
|
|
|
|
/*
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
If thread which holds conflicting lock is waiting on table-level
|
|
|
|
lock or some other non-MDL resource we might need to wake it up
|
|
|
|
by calling code outside of MDL.
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
*/
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
mysql_notify_thread_having_shared_lock(thd, conflicting_thd,
|
|
|
|
conflicting_ctx->get_needs_thr_lock_abort());
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
Auxiliary method for acquiring an exclusive lock.
|
|
|
|
|
|
|
|
@param mdl_request Request for the lock to be acqured.
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
@note Should not be used outside of MDL subsystem. Instead one
|
|
|
|
should call acquire_lock() or acquire_locks()
|
|
|
|
methods which ensure that conditions for deadlock-free
|
|
|
|
lock acquisition are fulfilled.
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
|
|
|
@retval FALSE Success
|
|
|
|
@retval TRUE Failure
|
|
|
|
*/
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
bool MDL_context::acquire_lock_impl(MDL_request *mdl_request)
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
{
|
|
|
|
MDL_lock *lock;
|
|
|
|
MDL_ticket *ticket;
|
|
|
|
bool not_used;
|
|
|
|
st_my_thread_var *mysys_var= my_thread_var;
|
|
|
|
MDL_key *key= &mdl_request->key;
|
|
|
|
|
2010-02-02 00:22:16 +01:00
|
|
|
mysql_mutex_assert_not_owner(&LOCK_open);
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
DBUG_ASSERT(mdl_request->ticket == NULL);
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
/* Don't take chances in production. */
|
|
|
|
mdl_request->ticket= NULL;
|
|
|
|
|
|
|
|
/*
|
|
|
|
Check whether the context already holds an exclusive lock on the object,
|
|
|
|
and if so, grant the request.
|
|
|
|
*/
|
|
|
|
if ((ticket= find_ticket(mdl_request, ¬_used)))
|
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
DBUG_ASSERT(ticket->m_lock);
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
mdl_request->ticket= ticket;
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
DBUG_ASSERT(mdl_request->type < MDL_SHARED_NO_WRITE ||
|
|
|
|
is_lock_owner(MDL_key::GLOBAL, "", "", MDL_INTENTION_EXCLUSIVE));
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
|
|
|
/* Early allocation: ticket will be needed in any case. */
|
|
|
|
if (!(ticket= MDL_ticket::create(this, mdl_request->type)))
|
|
|
|
return TRUE;
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
/* The below call implicitly locks MDL_lock::m_rwlock on success. */
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
if (!(lock= mdl_locks.find_or_insert(key)))
|
|
|
|
{
|
|
|
|
MDL_ticket::destroy(ticket);
|
|
|
|
return TRUE;
|
|
|
|
}
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
ticket->m_lock= lock;
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
lock->m_waiting.add_ticket(ticket);
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
while (!lock->can_grant_lock(mdl_request->type, this))
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
wait_reset();
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
if (ticket->is_upgradable_or_exclusive())
|
|
|
|
lock->notify_shared_locks(this);
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
rw_unlock(&lock->m_rwlock);
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
set_deadlock_weight(mdl_request->get_deadlock_weight());
|
|
|
|
will_wait_for(ticket);
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
|
|
|
/* There is a shared or exclusive lock on the object. */
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
DEBUG_SYNC(m_thd, "mdl_acquire_lock_wait");
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
bool is_deadlock= (find_deadlock() || timed_wait(1) == VICTIM_WAKE_UP);
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
stop_waiting();
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
if (is_deadlock || mysys_var->abort)
|
|
|
|
{
|
|
|
|
lock->remove_ticket(&MDL_lock::m_waiting, ticket);
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
MDL_ticket::destroy(ticket);
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
if (is_deadlock)
|
|
|
|
my_error(ER_LOCK_DEADLOCK, MYF(0));
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
return TRUE;
|
|
|
|
}
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
rw_wrlock(&lock->m_rwlock);
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
}
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
lock->m_waiting.remove_ticket(ticket);
|
|
|
|
lock->m_granted.add_ticket(ticket);
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
if (ticket->get_type() == MDL_EXCLUSIVE && lock->cached_object)
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
(*lock->cached_object_release_hook)(lock->cached_object);
|
|
|
|
lock->cached_object= NULL;
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
rw_unlock(&lock->m_rwlock);
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
|
|
|
|
m_tickets.push_front(ticket);
|
|
|
|
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
mdl_request->ticket= ticket;
|
|
|
|
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
|
|
|
extern "C" int mdl_request_ptr_cmp(const void* ptr1, const void* ptr2)
|
Backport of:
----------------------------------------------------------
revno: 2617.69.20
committer: Konstantin Osipov <kostja@sun.com>
branch nick: 5.4-4284-1-assert
timestamp: Thu 2009-08-13 18:29:55 +0400
message:
WL#4284 "Transactional DDL locking"
A review fix.
Since WL#4284 implementation separated MDL_request and MDL_ticket,
MDL_request becamse a utility object necessary only to get a ticket.
Store it by-value in TABLE_LIST with the intent to merge
MDL_request::key with table_list->table_name and table_list->db
in future.
Change the MDL subsystem to not require MDL_requests to
stay around till close_thread_tables().
Remove the list of requests from the MDL context.
Requests for shared metadata locks acquired in open_tables()
are only used as a list in recover_from_failed_open_table_attempt(),
which calls mdl_context.wait_for_locks() for this list.
To keep such list for recover_from_failed_open_table_attempt(),
introduce a context class (Open_table_context), that collects
all requests.
A lot of minor cleanups and simplications that became possible
with this change.
sql/event_db_repository.cc:
Remove alloc_mdl_requests(). Now MDL_request instance is a member
of TABLE_LIST, and init_one_table() initializes it.
sql/ha_ndbcluster_binlog.cc:
Remove now unnecessary declaration and initialization
of binlog_mdl_request.
sql/lock.cc:
No need to allocate MDL requests in lock_table_names() now.
sql/log.cc:
Use init_one_table() method, remove alloc_mdl_requests(),
which is now unnecessary.
sql/log_event.cc:
No need to allocate mdl_request separately now.
Use init_one_table() method.
sql/log_event_old.cc:
Update to the new signature of close_tables_for_reopen().
sql/mdl.cc:
Update try_acquire_exclusive_lock() to be more easy to use.
Function lock_table_name_if_not_cached() has been removed.
Make acquire_shared_lock() signature consistent with
try_acquire_exclusive_lock() signature.
Remove methods that are no longer used.
Update comments.
sql/mdl.h:
Implement an assignment operator that doesn't
copy MDL_key (MDL_key::operator= is private and
should remain private).
This is a hack to work-around assignment of TABLE_LIST
by value in several places. Such assignments violate
encapsulation, since only perform a shallow copy.
In most cases these assignments are a hack on their own.
sql/mysql_priv.h:
Update signatures of close_thread_tables() and close_tables_for_reopen().
sql/sp.cc:
Allocate TABLE_LIST in thd->mem_root.
Use init_one_table().
sql/sp_head.cc:
Use init_one_table().
Remove thd->locked_tables_root, it's no longer needed.
sql/sql_acl.cc:
Use init_mdl_requests() and init_one_table().
sql/sql_base.cc:
Update to new signatures of try_acquire_shared_lock() and
try_acquire_exclusive_lock().
Remove lock_table_name_if_not_cached().
Fix a bug in open_ltable() that would not return ER_LOCK_DEADLOCK
in case of a failed lock_tables() and a multi-statement
transaction.
Fix a bug in open_and_lock_tables_derived() that would
not return ER_LOCK_DEADLOCK in case of a multi-statement
transaction and a failure of lock_tables().
Move assignment of enum_open_table_action to a method of Open_table_context, a new class that maintains information
for backoff actions.
Minor rearrangements of the code.
Remove alloc_mdl_requests() in functions that work with system
tables: instead the patch ensures that callers always initialize
TABLE_LIST argument.
sql/sql_class.cc:
THD::locked_tables_root is no more.
sql/sql_class.h:
THD::locked_tables_root is no more.
Add a declaration for Open_table_context class.
sql/sql_delete.cc:
Update to use the simplified MDL API.
sql/sql_handler.cc:
TABLE_LIST::mdl_request is stored by-value now.
Ensure that mdl_request.ticket is NULL for every request
that is passed into MDL, to satisfy MDL asserts.
@ sql/sql_help.cc
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_help.cc:
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_insert.cc:
Remove assignment by-value of TABLE_LIST in
TABLEOP_HOOKS. We can't carry over a granted
MDL ticket from one table list to another.
sql/sql_parse.cc:
Change alloc_mdl_requests() -> init_mdl_requests().
@todo We can remove init_mdl_requests() altogether
in some places: all places that call add_table_to_list()
already have mdl requests initialized.
sql/sql_plugin.cc:
Use init_one_table().
THD::locked_tables_root is no more.
sql/sql_servers.cc:
Use init_one_table().
sql/sql_show.cc:
Update acquire_high_priority_shared_lock() to use
TABLE_LIST::mdl_request rather than allocate an own.
Fix get_trigger_table_impl() to use init_one_table(),
check for out of memory, follow the coding style.
sql/sql_table.cc:
Update to work with TABLE_LIST::mdl_request by-value.
Remove lock_table_name_if_not_cached().
The code that used to delegate to it is quite simple and
concise without it now.
sql/sql_udf.cc:
Use init_one_table().
sql/sql_update.cc:
Update to use the new signature of close_tables_for_reopen().
sql/table.cc:
Move re-setting of mdl_requests for prepared statements
and stored procedures from close_thread_tables() to
reinit_stmt_before_use().
Change alloc_mdl_requests() to init_mdl_requests().
init_mdl_requests() is a hack that can't be deleted
until we don't have a list-aware TABLE_LIST constructor.
Hopefully its use will be minimal
sql/table.h:
Change alloc_mdl_requests() to init_mdl_requests()
TABLE_LIST::mdl_request is stored by value.
sql/tztime.cc:
We no longer initialize mdl requests in open_system_tables_for*()
functions. Move this initialization closer to initialization
of the rest of TABLE_LIST members.
storage/myisammrg/ha_myisammrg.cc:
Simplify mdl_request initialization.
2009-12-08 10:57:07 +01:00
|
|
|
{
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
MDL_request *req1= *(MDL_request**)ptr1;
|
|
|
|
MDL_request *req2= *(MDL_request**)ptr2;
|
|
|
|
return req1->key.cmp(&req2->key);
|
Backport of:
----------------------------------------------------------
revno: 2617.69.20
committer: Konstantin Osipov <kostja@sun.com>
branch nick: 5.4-4284-1-assert
timestamp: Thu 2009-08-13 18:29:55 +0400
message:
WL#4284 "Transactional DDL locking"
A review fix.
Since WL#4284 implementation separated MDL_request and MDL_ticket,
MDL_request becamse a utility object necessary only to get a ticket.
Store it by-value in TABLE_LIST with the intent to merge
MDL_request::key with table_list->table_name and table_list->db
in future.
Change the MDL subsystem to not require MDL_requests to
stay around till close_thread_tables().
Remove the list of requests from the MDL context.
Requests for shared metadata locks acquired in open_tables()
are only used as a list in recover_from_failed_open_table_attempt(),
which calls mdl_context.wait_for_locks() for this list.
To keep such list for recover_from_failed_open_table_attempt(),
introduce a context class (Open_table_context), that collects
all requests.
A lot of minor cleanups and simplications that became possible
with this change.
sql/event_db_repository.cc:
Remove alloc_mdl_requests(). Now MDL_request instance is a member
of TABLE_LIST, and init_one_table() initializes it.
sql/ha_ndbcluster_binlog.cc:
Remove now unnecessary declaration and initialization
of binlog_mdl_request.
sql/lock.cc:
No need to allocate MDL requests in lock_table_names() now.
sql/log.cc:
Use init_one_table() method, remove alloc_mdl_requests(),
which is now unnecessary.
sql/log_event.cc:
No need to allocate mdl_request separately now.
Use init_one_table() method.
sql/log_event_old.cc:
Update to the new signature of close_tables_for_reopen().
sql/mdl.cc:
Update try_acquire_exclusive_lock() to be more easy to use.
Function lock_table_name_if_not_cached() has been removed.
Make acquire_shared_lock() signature consistent with
try_acquire_exclusive_lock() signature.
Remove methods that are no longer used.
Update comments.
sql/mdl.h:
Implement an assignment operator that doesn't
copy MDL_key (MDL_key::operator= is private and
should remain private).
This is a hack to work-around assignment of TABLE_LIST
by value in several places. Such assignments violate
encapsulation, since only perform a shallow copy.
In most cases these assignments are a hack on their own.
sql/mysql_priv.h:
Update signatures of close_thread_tables() and close_tables_for_reopen().
sql/sp.cc:
Allocate TABLE_LIST in thd->mem_root.
Use init_one_table().
sql/sp_head.cc:
Use init_one_table().
Remove thd->locked_tables_root, it's no longer needed.
sql/sql_acl.cc:
Use init_mdl_requests() and init_one_table().
sql/sql_base.cc:
Update to new signatures of try_acquire_shared_lock() and
try_acquire_exclusive_lock().
Remove lock_table_name_if_not_cached().
Fix a bug in open_ltable() that would not return ER_LOCK_DEADLOCK
in case of a failed lock_tables() and a multi-statement
transaction.
Fix a bug in open_and_lock_tables_derived() that would
not return ER_LOCK_DEADLOCK in case of a multi-statement
transaction and a failure of lock_tables().
Move assignment of enum_open_table_action to a method of Open_table_context, a new class that maintains information
for backoff actions.
Minor rearrangements of the code.
Remove alloc_mdl_requests() in functions that work with system
tables: instead the patch ensures that callers always initialize
TABLE_LIST argument.
sql/sql_class.cc:
THD::locked_tables_root is no more.
sql/sql_class.h:
THD::locked_tables_root is no more.
Add a declaration for Open_table_context class.
sql/sql_delete.cc:
Update to use the simplified MDL API.
sql/sql_handler.cc:
TABLE_LIST::mdl_request is stored by-value now.
Ensure that mdl_request.ticket is NULL for every request
that is passed into MDL, to satisfy MDL asserts.
@ sql/sql_help.cc
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_help.cc:
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_insert.cc:
Remove assignment by-value of TABLE_LIST in
TABLEOP_HOOKS. We can't carry over a granted
MDL ticket from one table list to another.
sql/sql_parse.cc:
Change alloc_mdl_requests() -> init_mdl_requests().
@todo We can remove init_mdl_requests() altogether
in some places: all places that call add_table_to_list()
already have mdl requests initialized.
sql/sql_plugin.cc:
Use init_one_table().
THD::locked_tables_root is no more.
sql/sql_servers.cc:
Use init_one_table().
sql/sql_show.cc:
Update acquire_high_priority_shared_lock() to use
TABLE_LIST::mdl_request rather than allocate an own.
Fix get_trigger_table_impl() to use init_one_table(),
check for out of memory, follow the coding style.
sql/sql_table.cc:
Update to work with TABLE_LIST::mdl_request by-value.
Remove lock_table_name_if_not_cached().
The code that used to delegate to it is quite simple and
concise without it now.
sql/sql_udf.cc:
Use init_one_table().
sql/sql_update.cc:
Update to use the new signature of close_tables_for_reopen().
sql/table.cc:
Move re-setting of mdl_requests for prepared statements
and stored procedures from close_thread_tables() to
reinit_stmt_before_use().
Change alloc_mdl_requests() to init_mdl_requests().
init_mdl_requests() is a hack that can't be deleted
until we don't have a list-aware TABLE_LIST constructor.
Hopefully its use will be minimal
sql/table.h:
Change alloc_mdl_requests() to init_mdl_requests()
TABLE_LIST::mdl_request is stored by value.
sql/tztime.cc:
We no longer initialize mdl requests in open_system_tables_for*()
functions. Move this initialization closer to initialization
of the rest of TABLE_LIST members.
storage/myisammrg/ha_myisammrg.cc:
Simplify mdl_request initialization.
2009-12-08 10:57:07 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
/**
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
Acquire exclusive locks. There must be no granted locks in the
|
2009-12-04 00:34:19 +01:00
|
|
|
context.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
This is a replacement of lock_table_names(). It is used in
|
|
|
|
RENAME, DROP and other DDL SQL statements.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
@param mdl_requests List of requests for locks to be acquired.
|
|
|
|
|
|
|
|
@note The list of requests should not contain non-exclusive lock requests.
|
|
|
|
There should not be any acquired locks in the context.
|
|
|
|
|
|
|
|
@note Assumes that one already owns global intention exclusive lock.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
@retval FALSE Success
|
|
|
|
@retval TRUE Failure
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
*/
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
bool MDL_context::acquire_locks(MDL_request_list *mdl_requests)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
Backport of:
----------------------------------------------------------
revno: 2617.69.20
committer: Konstantin Osipov <kostja@sun.com>
branch nick: 5.4-4284-1-assert
timestamp: Thu 2009-08-13 18:29:55 +0400
message:
WL#4284 "Transactional DDL locking"
A review fix.
Since WL#4284 implementation separated MDL_request and MDL_ticket,
MDL_request becamse a utility object necessary only to get a ticket.
Store it by-value in TABLE_LIST with the intent to merge
MDL_request::key with table_list->table_name and table_list->db
in future.
Change the MDL subsystem to not require MDL_requests to
stay around till close_thread_tables().
Remove the list of requests from the MDL context.
Requests for shared metadata locks acquired in open_tables()
are only used as a list in recover_from_failed_open_table_attempt(),
which calls mdl_context.wait_for_locks() for this list.
To keep such list for recover_from_failed_open_table_attempt(),
introduce a context class (Open_table_context), that collects
all requests.
A lot of minor cleanups and simplications that became possible
with this change.
sql/event_db_repository.cc:
Remove alloc_mdl_requests(). Now MDL_request instance is a member
of TABLE_LIST, and init_one_table() initializes it.
sql/ha_ndbcluster_binlog.cc:
Remove now unnecessary declaration and initialization
of binlog_mdl_request.
sql/lock.cc:
No need to allocate MDL requests in lock_table_names() now.
sql/log.cc:
Use init_one_table() method, remove alloc_mdl_requests(),
which is now unnecessary.
sql/log_event.cc:
No need to allocate mdl_request separately now.
Use init_one_table() method.
sql/log_event_old.cc:
Update to the new signature of close_tables_for_reopen().
sql/mdl.cc:
Update try_acquire_exclusive_lock() to be more easy to use.
Function lock_table_name_if_not_cached() has been removed.
Make acquire_shared_lock() signature consistent with
try_acquire_exclusive_lock() signature.
Remove methods that are no longer used.
Update comments.
sql/mdl.h:
Implement an assignment operator that doesn't
copy MDL_key (MDL_key::operator= is private and
should remain private).
This is a hack to work-around assignment of TABLE_LIST
by value in several places. Such assignments violate
encapsulation, since only perform a shallow copy.
In most cases these assignments are a hack on their own.
sql/mysql_priv.h:
Update signatures of close_thread_tables() and close_tables_for_reopen().
sql/sp.cc:
Allocate TABLE_LIST in thd->mem_root.
Use init_one_table().
sql/sp_head.cc:
Use init_one_table().
Remove thd->locked_tables_root, it's no longer needed.
sql/sql_acl.cc:
Use init_mdl_requests() and init_one_table().
sql/sql_base.cc:
Update to new signatures of try_acquire_shared_lock() and
try_acquire_exclusive_lock().
Remove lock_table_name_if_not_cached().
Fix a bug in open_ltable() that would not return ER_LOCK_DEADLOCK
in case of a failed lock_tables() and a multi-statement
transaction.
Fix a bug in open_and_lock_tables_derived() that would
not return ER_LOCK_DEADLOCK in case of a multi-statement
transaction and a failure of lock_tables().
Move assignment of enum_open_table_action to a method of Open_table_context, a new class that maintains information
for backoff actions.
Minor rearrangements of the code.
Remove alloc_mdl_requests() in functions that work with system
tables: instead the patch ensures that callers always initialize
TABLE_LIST argument.
sql/sql_class.cc:
THD::locked_tables_root is no more.
sql/sql_class.h:
THD::locked_tables_root is no more.
Add a declaration for Open_table_context class.
sql/sql_delete.cc:
Update to use the simplified MDL API.
sql/sql_handler.cc:
TABLE_LIST::mdl_request is stored by-value now.
Ensure that mdl_request.ticket is NULL for every request
that is passed into MDL, to satisfy MDL asserts.
@ sql/sql_help.cc
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_help.cc:
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_insert.cc:
Remove assignment by-value of TABLE_LIST in
TABLEOP_HOOKS. We can't carry over a granted
MDL ticket from one table list to another.
sql/sql_parse.cc:
Change alloc_mdl_requests() -> init_mdl_requests().
@todo We can remove init_mdl_requests() altogether
in some places: all places that call add_table_to_list()
already have mdl requests initialized.
sql/sql_plugin.cc:
Use init_one_table().
THD::locked_tables_root is no more.
sql/sql_servers.cc:
Use init_one_table().
sql/sql_show.cc:
Update acquire_high_priority_shared_lock() to use
TABLE_LIST::mdl_request rather than allocate an own.
Fix get_trigger_table_impl() to use init_one_table(),
check for out of memory, follow the coding style.
sql/sql_table.cc:
Update to work with TABLE_LIST::mdl_request by-value.
Remove lock_table_name_if_not_cached().
The code that used to delegate to it is quite simple and
concise without it now.
sql/sql_udf.cc:
Use init_one_table().
sql/sql_update.cc:
Update to use the new signature of close_tables_for_reopen().
sql/table.cc:
Move re-setting of mdl_requests for prepared statements
and stored procedures from close_thread_tables() to
reinit_stmt_before_use().
Change alloc_mdl_requests() to init_mdl_requests().
init_mdl_requests() is a hack that can't be deleted
until we don't have a list-aware TABLE_LIST constructor.
Hopefully its use will be minimal
sql/table.h:
Change alloc_mdl_requests() to init_mdl_requests()
TABLE_LIST::mdl_request is stored by value.
sql/tztime.cc:
We no longer initialize mdl requests in open_system_tables_for*()
functions. Move this initialization closer to initialization
of the rest of TABLE_LIST members.
storage/myisammrg/ha_myisammrg.cc:
Simplify mdl_request initialization.
2009-12-08 10:57:07 +01:00
|
|
|
MDL_request_list::Iterator it(*mdl_requests);
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
MDL_request **sort_buf, **p_req;
|
2010-02-01 15:12:56 +01:00
|
|
|
ssize_t req_count= static_cast<ssize_t>(mdl_requests->elements());
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
|
|
|
|
if (req_count == 0)
|
|
|
|
return FALSE;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
/*
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
To reduce deadlocks, the server acquires all exclusive
|
|
|
|
locks at once. For shared locks, try_acquire_lock() is
|
|
|
|
used instead.
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
*/
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
DBUG_ASSERT(m_tickets.is_empty() || m_tickets.front() == m_trans_sentinel);
|
Backport of:
----------------------------------------------------------
revno: 2617.69.20
committer: Konstantin Osipov <kostja@sun.com>
branch nick: 5.4-4284-1-assert
timestamp: Thu 2009-08-13 18:29:55 +0400
message:
WL#4284 "Transactional DDL locking"
A review fix.
Since WL#4284 implementation separated MDL_request and MDL_ticket,
MDL_request becamse a utility object necessary only to get a ticket.
Store it by-value in TABLE_LIST with the intent to merge
MDL_request::key with table_list->table_name and table_list->db
in future.
Change the MDL subsystem to not require MDL_requests to
stay around till close_thread_tables().
Remove the list of requests from the MDL context.
Requests for shared metadata locks acquired in open_tables()
are only used as a list in recover_from_failed_open_table_attempt(),
which calls mdl_context.wait_for_locks() for this list.
To keep such list for recover_from_failed_open_table_attempt(),
introduce a context class (Open_table_context), that collects
all requests.
A lot of minor cleanups and simplications that became possible
with this change.
sql/event_db_repository.cc:
Remove alloc_mdl_requests(). Now MDL_request instance is a member
of TABLE_LIST, and init_one_table() initializes it.
sql/ha_ndbcluster_binlog.cc:
Remove now unnecessary declaration and initialization
of binlog_mdl_request.
sql/lock.cc:
No need to allocate MDL requests in lock_table_names() now.
sql/log.cc:
Use init_one_table() method, remove alloc_mdl_requests(),
which is now unnecessary.
sql/log_event.cc:
No need to allocate mdl_request separately now.
Use init_one_table() method.
sql/log_event_old.cc:
Update to the new signature of close_tables_for_reopen().
sql/mdl.cc:
Update try_acquire_exclusive_lock() to be more easy to use.
Function lock_table_name_if_not_cached() has been removed.
Make acquire_shared_lock() signature consistent with
try_acquire_exclusive_lock() signature.
Remove methods that are no longer used.
Update comments.
sql/mdl.h:
Implement an assignment operator that doesn't
copy MDL_key (MDL_key::operator= is private and
should remain private).
This is a hack to work-around assignment of TABLE_LIST
by value in several places. Such assignments violate
encapsulation, since only perform a shallow copy.
In most cases these assignments are a hack on their own.
sql/mysql_priv.h:
Update signatures of close_thread_tables() and close_tables_for_reopen().
sql/sp.cc:
Allocate TABLE_LIST in thd->mem_root.
Use init_one_table().
sql/sp_head.cc:
Use init_one_table().
Remove thd->locked_tables_root, it's no longer needed.
sql/sql_acl.cc:
Use init_mdl_requests() and init_one_table().
sql/sql_base.cc:
Update to new signatures of try_acquire_shared_lock() and
try_acquire_exclusive_lock().
Remove lock_table_name_if_not_cached().
Fix a bug in open_ltable() that would not return ER_LOCK_DEADLOCK
in case of a failed lock_tables() and a multi-statement
transaction.
Fix a bug in open_and_lock_tables_derived() that would
not return ER_LOCK_DEADLOCK in case of a multi-statement
transaction and a failure of lock_tables().
Move assignment of enum_open_table_action to a method of Open_table_context, a new class that maintains information
for backoff actions.
Minor rearrangements of the code.
Remove alloc_mdl_requests() in functions that work with system
tables: instead the patch ensures that callers always initialize
TABLE_LIST argument.
sql/sql_class.cc:
THD::locked_tables_root is no more.
sql/sql_class.h:
THD::locked_tables_root is no more.
Add a declaration for Open_table_context class.
sql/sql_delete.cc:
Update to use the simplified MDL API.
sql/sql_handler.cc:
TABLE_LIST::mdl_request is stored by-value now.
Ensure that mdl_request.ticket is NULL for every request
that is passed into MDL, to satisfy MDL asserts.
@ sql/sql_help.cc
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_help.cc:
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_insert.cc:
Remove assignment by-value of TABLE_LIST in
TABLEOP_HOOKS. We can't carry over a granted
MDL ticket from one table list to another.
sql/sql_parse.cc:
Change alloc_mdl_requests() -> init_mdl_requests().
@todo We can remove init_mdl_requests() altogether
in some places: all places that call add_table_to_list()
already have mdl requests initialized.
sql/sql_plugin.cc:
Use init_one_table().
THD::locked_tables_root is no more.
sql/sql_servers.cc:
Use init_one_table().
sql/sql_show.cc:
Update acquire_high_priority_shared_lock() to use
TABLE_LIST::mdl_request rather than allocate an own.
Fix get_trigger_table_impl() to use init_one_table(),
check for out of memory, follow the coding style.
sql/sql_table.cc:
Update to work with TABLE_LIST::mdl_request by-value.
Remove lock_table_name_if_not_cached().
The code that used to delegate to it is quite simple and
concise without it now.
sql/sql_udf.cc:
Use init_one_table().
sql/sql_update.cc:
Update to use the new signature of close_tables_for_reopen().
sql/table.cc:
Move re-setting of mdl_requests for prepared statements
and stored procedures from close_thread_tables() to
reinit_stmt_before_use().
Change alloc_mdl_requests() to init_mdl_requests().
init_mdl_requests() is a hack that can't be deleted
until we don't have a list-aware TABLE_LIST constructor.
Hopefully its use will be minimal
sql/table.h:
Change alloc_mdl_requests() to init_mdl_requests()
TABLE_LIST::mdl_request is stored by value.
sql/tztime.cc:
We no longer initialize mdl requests in open_system_tables_for*()
functions. Move this initialization closer to initialization
of the rest of TABLE_LIST members.
storage/myisammrg/ha_myisammrg.cc:
Simplify mdl_request initialization.
2009-12-08 10:57:07 +01:00
|
|
|
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
/* Sort requests according to MDL_key. */
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
if (! (sort_buf= (MDL_request **)my_malloc(req_count *
|
|
|
|
sizeof(MDL_request*),
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
MYF(MY_WME))))
|
|
|
|
return TRUE;
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
for (p_req= sort_buf; p_req < sort_buf + req_count; p_req++)
|
|
|
|
*p_req= it++;
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
my_qsort(sort_buf, req_count, sizeof(MDL_request*),
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
mdl_request_ptr_cmp);
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
for (p_req= sort_buf; p_req < sort_buf + req_count; p_req++)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
if (acquire_lock_impl(*p_req))
|
2009-12-02 17:15:40 +01:00
|
|
|
goto err;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
my_free(sort_buf, MYF(0));
|
2009-11-30 23:33:22 +01:00
|
|
|
return FALSE;
|
2009-12-02 17:15:40 +01:00
|
|
|
|
|
|
|
err:
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
/* Release locks we have managed to acquire so far. */
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
for (req_count= p_req - sort_buf, p_req= sort_buf;
|
|
|
|
p_req < sort_buf + req_count; p_req++)
|
2009-12-02 17:15:40 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
release_lock((*p_req)->ticket);
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
/* Reset lock request back to its initial state. */
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
(*p_req)->ticket= NULL;
|
2009-12-02 17:15:40 +01:00
|
|
|
}
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
my_free(sort_buf, MYF(0));
|
2009-12-02 17:15:40 +01:00
|
|
|
return TRUE;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
2009-12-04 00:34:19 +01:00
|
|
|
Upgrade a shared metadata lock to exclusive.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
Used in ALTER TABLE, when a copy of the table with the
|
|
|
|
new definition has been constructed.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
@note In case of failure to upgrade lock (e.g. because upgrader
|
|
|
|
was killed) leaves lock in its original state (locked in
|
|
|
|
shared mode).
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
@note There can be only one upgrader for a lock or we will have deadlock.
|
|
|
|
This invariant is ensured by code outside of metadata subsystem usually
|
|
|
|
by obtaining some sort of exclusive table-level lock (e.g. TL_WRITE,
|
|
|
|
TL_WRITE_ALLOW_READ) before performing upgrade of metadata lock.
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
@retval FALSE Success
|
|
|
|
@retval TRUE Failure (thread was killed)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
*/
|
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
bool
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
MDL_context::upgrade_shared_lock_to_exclusive(MDL_ticket *mdl_ticket)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
MDL_request mdl_xlock_request;
|
|
|
|
MDL_ticket *mdl_svp= mdl_savepoint();
|
|
|
|
bool is_new_ticket;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
DBUG_ENTER("MDL_ticket::upgrade_shared_lock_to_exclusive");
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
DEBUG_SYNC(get_thd(), "mdl_upgrade_shared_lock_to_exclusive");
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
/*
|
|
|
|
Do nothing if already upgraded. Used when we FLUSH TABLE under
|
|
|
|
LOCK TABLES and a table is listed twice in LOCK TABLES list.
|
|
|
|
*/
|
|
|
|
if (mdl_ticket->m_type == MDL_EXCLUSIVE)
|
2009-11-30 23:39:13 +01:00
|
|
|
DBUG_RETURN(FALSE);
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
/* Only allow upgrades from MDL_SHARED_NO_WRITE/NO_READ_WRITE */
|
|
|
|
DBUG_ASSERT(mdl_ticket->m_type == MDL_SHARED_NO_WRITE ||
|
|
|
|
mdl_ticket->m_type == MDL_SHARED_NO_READ_WRITE);
|
2009-12-09 10:44:01 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
mdl_xlock_request.init(&mdl_ticket->m_lock->key, MDL_EXCLUSIVE);
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
if (acquire_lock_impl(&mdl_xlock_request))
|
2009-12-30 18:53:30 +01:00
|
|
|
DBUG_RETURN(TRUE);
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
is_new_ticket= ! has_lock(mdl_svp, mdl_xlock_request.ticket);
|
2009-12-30 18:53:30 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
/* Merge the acquired and the original lock. @todo: move to a method. */
|
|
|
|
rw_wrlock(&mdl_ticket->m_lock->m_rwlock);
|
|
|
|
if (is_new_ticket)
|
|
|
|
mdl_ticket->m_lock->m_granted.remove_ticket(mdl_xlock_request.ticket);
|
|
|
|
/*
|
|
|
|
Set the new type of lock in the ticket. To update state of
|
|
|
|
MDL_lock object correctly we need to temporarily exclude
|
|
|
|
ticket from the granted queue and then include it back.
|
|
|
|
*/
|
|
|
|
mdl_ticket->m_lock->m_granted.remove_ticket(mdl_ticket);
|
|
|
|
mdl_ticket->m_type= MDL_EXCLUSIVE;
|
|
|
|
mdl_ticket->m_lock->m_granted.add_ticket(mdl_ticket);
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
rw_unlock(&mdl_ticket->m_lock->m_rwlock);
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
if (is_new_ticket)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
m_tickets.remove(mdl_xlock_request.ticket);
|
|
|
|
MDL_ticket::destroy(mdl_xlock_request.ticket);
|
|
|
|
}
|
2009-12-01 14:55:45 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
DBUG_RETURN(FALSE);
|
|
|
|
}
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
2009-12-30 18:53:30 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
bool MDL_lock::find_deadlock(MDL_ticket *waiting_ticket,
|
|
|
|
Deadlock_detection_context *deadlock_ctx)
|
|
|
|
{
|
|
|
|
MDL_ticket *ticket;
|
|
|
|
bool result= FALSE;
|
2009-11-30 23:39:13 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
rw_rdlock(&m_rwlock);
|
2009-12-10 11:53:20 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
Ticket_iterator granted_it(m_granted);
|
|
|
|
Ticket_iterator waiting_it(m_waiting);
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
while ((ticket= granted_it++))
|
|
|
|
{
|
|
|
|
if (ticket->is_incompatible_when_granted(waiting_ticket->get_type()) &&
|
|
|
|
ticket->get_ctx() != waiting_ticket->get_ctx() &&
|
|
|
|
ticket->get_ctx() == deadlock_ctx->start)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
result= TRUE;
|
|
|
|
goto end;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
while ((ticket= waiting_it++))
|
|
|
|
{
|
|
|
|
if (ticket->is_incompatible_when_waiting(waiting_ticket->get_type()) &&
|
|
|
|
ticket->get_ctx() != waiting_ticket->get_ctx() &&
|
|
|
|
ticket->get_ctx() == deadlock_ctx->start)
|
|
|
|
{
|
|
|
|
result= TRUE;
|
|
|
|
goto end;
|
|
|
|
}
|
|
|
|
}
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
granted_it.rewind();
|
|
|
|
while ((ticket= granted_it++))
|
|
|
|
{
|
|
|
|
if (ticket->is_incompatible_when_granted(waiting_ticket->get_type()) &&
|
|
|
|
ticket->get_ctx() != waiting_ticket->get_ctx() &&
|
|
|
|
ticket->get_ctx()->find_deadlock(deadlock_ctx))
|
|
|
|
{
|
|
|
|
result= TRUE;
|
|
|
|
goto end;
|
|
|
|
}
|
|
|
|
}
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
waiting_it.rewind();
|
|
|
|
while ((ticket= waiting_it++))
|
|
|
|
{
|
|
|
|
if (ticket->is_incompatible_when_waiting(waiting_ticket->get_type()) &&
|
|
|
|
ticket->get_ctx() != waiting_ticket->get_ctx() &&
|
|
|
|
ticket->get_ctx()->find_deadlock(deadlock_ctx))
|
|
|
|
{
|
|
|
|
result= TRUE;
|
|
|
|
goto end;
|
|
|
|
}
|
|
|
|
}
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
end:
|
|
|
|
rw_unlock(&m_rwlock);
|
|
|
|
return result;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
bool MDL_context::find_deadlock(Deadlock_detection_context *deadlock_ctx)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
bool result= FALSE;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
rw_rdlock(&m_waiting_for_lock);
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
if (m_waiting_for)
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
QQ: should we rather be checking for NO_WAKE_UP ?
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
We want to do check signal only when m_waiting_for is set
|
|
|
|
to avoid reading left-overs from previous kills.
|
|
|
|
*/
|
|
|
|
if (peek_signal() != VICTIM_WAKE_UP)
|
|
|
|
{
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
if (++deadlock_ctx->current_search_depth >
|
|
|
|
deadlock_ctx->MAX_SEARCH_DEPTH)
|
|
|
|
result= TRUE;
|
|
|
|
else
|
|
|
|
result= m_waiting_for->m_lock->find_deadlock(m_waiting_for,
|
|
|
|
deadlock_ctx);
|
|
|
|
--deadlock_ctx->current_search_depth;
|
|
|
|
}
|
|
|
|
}
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
if (result)
|
|
|
|
{
|
|
|
|
if (! deadlock_ctx->victim)
|
|
|
|
deadlock_ctx->victim= this;
|
|
|
|
else if (deadlock_ctx->victim->m_deadlock_weight >= m_deadlock_weight)
|
|
|
|
{
|
|
|
|
rw_unlock(&deadlock_ctx->victim->m_waiting_for_lock);
|
|
|
|
deadlock_ctx->victim= this;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
rw_unlock(&m_waiting_for_lock);
|
|
|
|
}
|
|
|
|
else
|
|
|
|
rw_unlock(&m_waiting_for_lock);
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
return result;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
bool MDL_context::find_deadlock()
|
2009-12-30 18:53:30 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
Deadlock_detection_context deadlock_ctx(this);
|
2009-12-30 18:53:30 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
while (1)
|
2009-12-30 18:53:30 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
if (! find_deadlock(&deadlock_ctx))
|
|
|
|
{
|
|
|
|
/* No deadlocks are found! */
|
|
|
|
break;
|
|
|
|
}
|
2009-12-30 18:53:30 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
if (deadlock_ctx.victim != this)
|
|
|
|
{
|
|
|
|
deadlock_ctx.victim->awake(VICTIM_WAKE_UP);
|
|
|
|
rw_unlock(&deadlock_ctx.victim->m_waiting_for_lock);
|
|
|
|
/*
|
|
|
|
After adding new arc to waiting graph we found that it participates
|
|
|
|
in some loop (i.e. there is a deadlock). We decided to destroy this
|
|
|
|
loop by removing some arc other than newly added. Since this doesn't
|
|
|
|
guarantee that all loops created by addition of this arc are
|
|
|
|
destroyed we have to repeat search.
|
|
|
|
*/
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
DBUG_ASSERT(&deadlock_ctx.victim->m_waiting_for_lock == &m_waiting_for_lock);
|
|
|
|
rw_unlock(&deadlock_ctx.victim->m_waiting_for_lock);
|
2009-12-30 18:53:30 +01:00
|
|
|
return TRUE;
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
}
|
2009-12-30 18:53:30 +01:00
|
|
|
}
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
/**
|
2009-12-04 00:34:19 +01:00
|
|
|
Wait until there will be no locks that conflict with lock requests
|
Backport of:
----------------------------------------------------------
revno: 2617.69.20
committer: Konstantin Osipov <kostja@sun.com>
branch nick: 5.4-4284-1-assert
timestamp: Thu 2009-08-13 18:29:55 +0400
message:
WL#4284 "Transactional DDL locking"
A review fix.
Since WL#4284 implementation separated MDL_request and MDL_ticket,
MDL_request becamse a utility object necessary only to get a ticket.
Store it by-value in TABLE_LIST with the intent to merge
MDL_request::key with table_list->table_name and table_list->db
in future.
Change the MDL subsystem to not require MDL_requests to
stay around till close_thread_tables().
Remove the list of requests from the MDL context.
Requests for shared metadata locks acquired in open_tables()
are only used as a list in recover_from_failed_open_table_attempt(),
which calls mdl_context.wait_for_locks() for this list.
To keep such list for recover_from_failed_open_table_attempt(),
introduce a context class (Open_table_context), that collects
all requests.
A lot of minor cleanups and simplications that became possible
with this change.
sql/event_db_repository.cc:
Remove alloc_mdl_requests(). Now MDL_request instance is a member
of TABLE_LIST, and init_one_table() initializes it.
sql/ha_ndbcluster_binlog.cc:
Remove now unnecessary declaration and initialization
of binlog_mdl_request.
sql/lock.cc:
No need to allocate MDL requests in lock_table_names() now.
sql/log.cc:
Use init_one_table() method, remove alloc_mdl_requests(),
which is now unnecessary.
sql/log_event.cc:
No need to allocate mdl_request separately now.
Use init_one_table() method.
sql/log_event_old.cc:
Update to the new signature of close_tables_for_reopen().
sql/mdl.cc:
Update try_acquire_exclusive_lock() to be more easy to use.
Function lock_table_name_if_not_cached() has been removed.
Make acquire_shared_lock() signature consistent with
try_acquire_exclusive_lock() signature.
Remove methods that are no longer used.
Update comments.
sql/mdl.h:
Implement an assignment operator that doesn't
copy MDL_key (MDL_key::operator= is private and
should remain private).
This is a hack to work-around assignment of TABLE_LIST
by value in several places. Such assignments violate
encapsulation, since only perform a shallow copy.
In most cases these assignments are a hack on their own.
sql/mysql_priv.h:
Update signatures of close_thread_tables() and close_tables_for_reopen().
sql/sp.cc:
Allocate TABLE_LIST in thd->mem_root.
Use init_one_table().
sql/sp_head.cc:
Use init_one_table().
Remove thd->locked_tables_root, it's no longer needed.
sql/sql_acl.cc:
Use init_mdl_requests() and init_one_table().
sql/sql_base.cc:
Update to new signatures of try_acquire_shared_lock() and
try_acquire_exclusive_lock().
Remove lock_table_name_if_not_cached().
Fix a bug in open_ltable() that would not return ER_LOCK_DEADLOCK
in case of a failed lock_tables() and a multi-statement
transaction.
Fix a bug in open_and_lock_tables_derived() that would
not return ER_LOCK_DEADLOCK in case of a multi-statement
transaction and a failure of lock_tables().
Move assignment of enum_open_table_action to a method of Open_table_context, a new class that maintains information
for backoff actions.
Minor rearrangements of the code.
Remove alloc_mdl_requests() in functions that work with system
tables: instead the patch ensures that callers always initialize
TABLE_LIST argument.
sql/sql_class.cc:
THD::locked_tables_root is no more.
sql/sql_class.h:
THD::locked_tables_root is no more.
Add a declaration for Open_table_context class.
sql/sql_delete.cc:
Update to use the simplified MDL API.
sql/sql_handler.cc:
TABLE_LIST::mdl_request is stored by-value now.
Ensure that mdl_request.ticket is NULL for every request
that is passed into MDL, to satisfy MDL asserts.
@ sql/sql_help.cc
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_help.cc:
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_insert.cc:
Remove assignment by-value of TABLE_LIST in
TABLEOP_HOOKS. We can't carry over a granted
MDL ticket from one table list to another.
sql/sql_parse.cc:
Change alloc_mdl_requests() -> init_mdl_requests().
@todo We can remove init_mdl_requests() altogether
in some places: all places that call add_table_to_list()
already have mdl requests initialized.
sql/sql_plugin.cc:
Use init_one_table().
THD::locked_tables_root is no more.
sql/sql_servers.cc:
Use init_one_table().
sql/sql_show.cc:
Update acquire_high_priority_shared_lock() to use
TABLE_LIST::mdl_request rather than allocate an own.
Fix get_trigger_table_impl() to use init_one_table(),
check for out of memory, follow the coding style.
sql/sql_table.cc:
Update to work with TABLE_LIST::mdl_request by-value.
Remove lock_table_name_if_not_cached().
The code that used to delegate to it is quite simple and
concise without it now.
sql/sql_udf.cc:
Use init_one_table().
sql/sql_update.cc:
Update to use the new signature of close_tables_for_reopen().
sql/table.cc:
Move re-setting of mdl_requests for prepared statements
and stored procedures from close_thread_tables() to
reinit_stmt_before_use().
Change alloc_mdl_requests() to init_mdl_requests().
init_mdl_requests() is a hack that can't be deleted
until we don't have a list-aware TABLE_LIST constructor.
Hopefully its use will be minimal
sql/table.h:
Change alloc_mdl_requests() to init_mdl_requests()
TABLE_LIST::mdl_request is stored by value.
sql/tztime.cc:
We no longer initialize mdl requests in open_system_tables_for*()
functions. Move this initialization closer to initialization
of the rest of TABLE_LIST members.
storage/myisammrg/ha_myisammrg.cc:
Simplify mdl_request initialization.
2009-12-08 10:57:07 +01:00
|
|
|
in the given list.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
This is a part of the locking protocol and must be used by the
|
|
|
|
acquirer of shared locks after a back-off.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
Does not acquire the locks!
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
@retval FALSE Success. One can try to obtain metadata locks.
|
2009-12-30 18:53:30 +01:00
|
|
|
@retval TRUE Failure (thread was killed or deadlock is possible).
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
*/
|
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
bool
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
MDL_context::wait_for_lock(MDL_request *mdl_request)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
2009-12-04 00:52:05 +01:00
|
|
|
MDL_lock *lock;
|
2009-12-02 17:31:57 +01:00
|
|
|
st_my_thread_var *mysys_var= my_thread_var;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2010-02-02 00:22:16 +01:00
|
|
|
mysql_mutex_assert_not_owner(&LOCK_open);
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
DBUG_ASSERT(mdl_request->ticket == NULL);
|
|
|
|
|
2009-12-02 17:31:57 +01:00
|
|
|
while (!mysys_var->abort)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
|
|
|
/*
|
|
|
|
We have to check if there are some HANDLERs open by this thread
|
|
|
|
which conflict with some pending exclusive locks. Otherwise we
|
|
|
|
might have a deadlock in situations when we are waiting for
|
|
|
|
pending writer to go away, which in its turn waits for HANDLER
|
|
|
|
open by our thread.
|
|
|
|
|
|
|
|
TODO: investigate situations in which we need to broadcast on
|
|
|
|
COND_mdl because of above scenario.
|
|
|
|
*/
|
2009-12-04 00:52:05 +01:00
|
|
|
mysql_ha_flush(m_thd);
|
2009-12-30 18:53:30 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
MDL_key *key= &mdl_request->key;
|
|
|
|
|
|
|
|
/* The below call implicitly locks MDL_lock::m_rwlock on success. */
|
|
|
|
if (! (lock= mdl_locks.find(key)))
|
|
|
|
return FALSE;
|
|
|
|
|
|
|
|
if (lock->can_grant_lock(mdl_request->type, this))
|
2009-12-30 18:53:30 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
rw_unlock(&lock->m_rwlock);
|
|
|
|
return FALSE;
|
2009-12-30 18:53:30 +01:00
|
|
|
}
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
MDL_ticket *pending_ticket;
|
|
|
|
if (! (pending_ticket= MDL_ticket::create(this, mdl_request->type)))
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
rw_unlock(&lock->m_rwlock);
|
|
|
|
return TRUE;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
|
|
|
|
pending_ticket->m_lock= lock;
|
|
|
|
|
|
|
|
lock->m_waiting.add_ticket(pending_ticket);
|
|
|
|
|
|
|
|
wait_reset();
|
|
|
|
rw_unlock(&lock->m_rwlock);
|
|
|
|
|
|
|
|
set_deadlock_weight(MDL_DEADLOCK_WEIGHT_DML);
|
|
|
|
will_wait_for(pending_ticket);
|
|
|
|
|
|
|
|
bool is_deadlock= (find_deadlock() || wait() == VICTIM_WAKE_UP);
|
|
|
|
|
|
|
|
stop_waiting();
|
|
|
|
|
|
|
|
lock->remove_ticket(&MDL_lock::m_waiting, pending_ticket);
|
|
|
|
MDL_ticket::destroy(pending_ticket);
|
|
|
|
if (is_deadlock)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
my_error(ER_LOCK_DEADLOCK, MYF(0));
|
|
|
|
return TRUE;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
|
|
|
}
|
2009-12-02 17:31:57 +01:00
|
|
|
return mysys_var->abort;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
Release lock.
|
|
|
|
|
|
|
|
@param ticket Ticket for lock to be released.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
*/
|
|
|
|
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
void MDL_context::release_lock(MDL_ticket *ticket)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
2009-12-04 00:52:05 +01:00
|
|
|
MDL_lock *lock= ticket->m_lock;
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
DBUG_ENTER("MDL_context::release_lock");
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
DBUG_PRINT("enter", ("db=%s name=%s", lock->key.db_name(),
|
2009-12-09 17:11:26 +01:00
|
|
|
lock->key.name()));
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
DBUG_ASSERT(this == ticket->get_ctx());
|
2010-02-02 00:22:16 +01:00
|
|
|
mysql_mutex_assert_not_owner(&LOCK_open);
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
if (ticket == m_trans_sentinel)
|
|
|
|
m_trans_sentinel= ++Ticket_list::Iterator(m_tickets, ticket);
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
lock->remove_ticket(&MDL_lock::m_granted, ticket);
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
m_tickets.remove(ticket);
|
|
|
|
MDL_ticket::destroy(ticket);
|
|
|
|
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
DBUG_VOID_RETURN;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
Release all locks associated with the context. If the sentinel
|
|
|
|
is not NULL, do not release locks stored in the list after and
|
|
|
|
including the sentinel.
|
|
|
|
|
|
|
|
Transactional locks are added to the beginning of the list, i.e.
|
|
|
|
stored in reverse temporal order. This allows to employ this
|
|
|
|
function to:
|
|
|
|
- back off in case of a lock conflict.
|
|
|
|
- release all locks in the end of a transaction
|
|
|
|
- rollback to a savepoint.
|
|
|
|
|
|
|
|
The sentinel semantics is used to support LOCK TABLES
|
|
|
|
mode and HANDLER statements: locks taken by these statements
|
|
|
|
survive COMMIT, ROLLBACK, ROLLBACK TO SAVEPOINT.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
*/
|
|
|
|
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
void MDL_context::release_locks_stored_before(MDL_ticket *sentinel)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
2009-12-04 00:52:05 +01:00
|
|
|
MDL_ticket *ticket;
|
|
|
|
Ticket_iterator it(m_tickets);
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
DBUG_ENTER("MDL_context::release_locks_stored_before");
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
if (m_tickets.is_empty())
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
DBUG_VOID_RETURN;
|
|
|
|
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
while ((ticket= it++) && ticket != sentinel)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
DBUG_PRINT("info", ("found lock to release ticket=%p", ticket));
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
release_lock(ticket);
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
/*
|
|
|
|
If all locks were released, then the sentinel was not present
|
|
|
|
in the list. It must never happen because the sentinel was
|
|
|
|
bogus, i.e. pointed to a ticket that no longer exists.
|
|
|
|
*/
|
|
|
|
DBUG_ASSERT(! m_tickets.is_empty() || sentinel == NULL);
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
DBUG_VOID_RETURN;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
2009-12-04 00:34:19 +01:00
|
|
|
Release all locks in the context which correspond to the same name/
|
Backport of:
----------------------------------------------------------
revno: 2617.69.20
committer: Konstantin Osipov <kostja@sun.com>
branch nick: 5.4-4284-1-assert
timestamp: Thu 2009-08-13 18:29:55 +0400
message:
WL#4284 "Transactional DDL locking"
A review fix.
Since WL#4284 implementation separated MDL_request and MDL_ticket,
MDL_request becamse a utility object necessary only to get a ticket.
Store it by-value in TABLE_LIST with the intent to merge
MDL_request::key with table_list->table_name and table_list->db
in future.
Change the MDL subsystem to not require MDL_requests to
stay around till close_thread_tables().
Remove the list of requests from the MDL context.
Requests for shared metadata locks acquired in open_tables()
are only used as a list in recover_from_failed_open_table_attempt(),
which calls mdl_context.wait_for_locks() for this list.
To keep such list for recover_from_failed_open_table_attempt(),
introduce a context class (Open_table_context), that collects
all requests.
A lot of minor cleanups and simplications that became possible
with this change.
sql/event_db_repository.cc:
Remove alloc_mdl_requests(). Now MDL_request instance is a member
of TABLE_LIST, and init_one_table() initializes it.
sql/ha_ndbcluster_binlog.cc:
Remove now unnecessary declaration and initialization
of binlog_mdl_request.
sql/lock.cc:
No need to allocate MDL requests in lock_table_names() now.
sql/log.cc:
Use init_one_table() method, remove alloc_mdl_requests(),
which is now unnecessary.
sql/log_event.cc:
No need to allocate mdl_request separately now.
Use init_one_table() method.
sql/log_event_old.cc:
Update to the new signature of close_tables_for_reopen().
sql/mdl.cc:
Update try_acquire_exclusive_lock() to be more easy to use.
Function lock_table_name_if_not_cached() has been removed.
Make acquire_shared_lock() signature consistent with
try_acquire_exclusive_lock() signature.
Remove methods that are no longer used.
Update comments.
sql/mdl.h:
Implement an assignment operator that doesn't
copy MDL_key (MDL_key::operator= is private and
should remain private).
This is a hack to work-around assignment of TABLE_LIST
by value in several places. Such assignments violate
encapsulation, since only perform a shallow copy.
In most cases these assignments are a hack on their own.
sql/mysql_priv.h:
Update signatures of close_thread_tables() and close_tables_for_reopen().
sql/sp.cc:
Allocate TABLE_LIST in thd->mem_root.
Use init_one_table().
sql/sp_head.cc:
Use init_one_table().
Remove thd->locked_tables_root, it's no longer needed.
sql/sql_acl.cc:
Use init_mdl_requests() and init_one_table().
sql/sql_base.cc:
Update to new signatures of try_acquire_shared_lock() and
try_acquire_exclusive_lock().
Remove lock_table_name_if_not_cached().
Fix a bug in open_ltable() that would not return ER_LOCK_DEADLOCK
in case of a failed lock_tables() and a multi-statement
transaction.
Fix a bug in open_and_lock_tables_derived() that would
not return ER_LOCK_DEADLOCK in case of a multi-statement
transaction and a failure of lock_tables().
Move assignment of enum_open_table_action to a method of Open_table_context, a new class that maintains information
for backoff actions.
Minor rearrangements of the code.
Remove alloc_mdl_requests() in functions that work with system
tables: instead the patch ensures that callers always initialize
TABLE_LIST argument.
sql/sql_class.cc:
THD::locked_tables_root is no more.
sql/sql_class.h:
THD::locked_tables_root is no more.
Add a declaration for Open_table_context class.
sql/sql_delete.cc:
Update to use the simplified MDL API.
sql/sql_handler.cc:
TABLE_LIST::mdl_request is stored by-value now.
Ensure that mdl_request.ticket is NULL for every request
that is passed into MDL, to satisfy MDL asserts.
@ sql/sql_help.cc
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_help.cc:
Function open_system_tables_for_read() no longer initializes
mdl_requests.
Move TABLE_LIST::mdl_request initialization closer to
TABLE_LIST initialization.
sql/sql_insert.cc:
Remove assignment by-value of TABLE_LIST in
TABLEOP_HOOKS. We can't carry over a granted
MDL ticket from one table list to another.
sql/sql_parse.cc:
Change alloc_mdl_requests() -> init_mdl_requests().
@todo We can remove init_mdl_requests() altogether
in some places: all places that call add_table_to_list()
already have mdl requests initialized.
sql/sql_plugin.cc:
Use init_one_table().
THD::locked_tables_root is no more.
sql/sql_servers.cc:
Use init_one_table().
sql/sql_show.cc:
Update acquire_high_priority_shared_lock() to use
TABLE_LIST::mdl_request rather than allocate an own.
Fix get_trigger_table_impl() to use init_one_table(),
check for out of memory, follow the coding style.
sql/sql_table.cc:
Update to work with TABLE_LIST::mdl_request by-value.
Remove lock_table_name_if_not_cached().
The code that used to delegate to it is quite simple and
concise without it now.
sql/sql_udf.cc:
Use init_one_table().
sql/sql_update.cc:
Update to use the new signature of close_tables_for_reopen().
sql/table.cc:
Move re-setting of mdl_requests for prepared statements
and stored procedures from close_thread_tables() to
reinit_stmt_before_use().
Change alloc_mdl_requests() to init_mdl_requests().
init_mdl_requests() is a hack that can't be deleted
until we don't have a list-aware TABLE_LIST constructor.
Hopefully its use will be minimal
sql/table.h:
Change alloc_mdl_requests() to init_mdl_requests()
TABLE_LIST::mdl_request is stored by value.
sql/tztime.cc:
We no longer initialize mdl requests in open_system_tables_for*()
functions. Move this initialization closer to initialization
of the rest of TABLE_LIST members.
storage/myisammrg/ha_myisammrg.cc:
Simplify mdl_request initialization.
2009-12-08 10:57:07 +01:00
|
|
|
object as this lock request.
|
2009-11-30 23:39:13 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
@param ticket One of the locks for the name/object for which all
|
|
|
|
locks should be released.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
*/
|
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
void MDL_context::release_all_locks_for_name(MDL_ticket *name)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
/* Use MDL_ticket::m_lock to identify other locks for the same object. */
|
2009-12-04 00:52:05 +01:00
|
|
|
MDL_lock *lock= name->m_lock;
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
|
|
|
|
/* Remove matching lock tickets from the context. */
|
2009-12-04 00:52:05 +01:00
|
|
|
MDL_ticket *ticket;
|
|
|
|
Ticket_iterator it_ticket(m_tickets);
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
while ((ticket= it_ticket++))
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
DBUG_ASSERT(ticket->m_lock);
|
2009-12-04 00:52:05 +01:00
|
|
|
/*
|
|
|
|
We rarely have more than one ticket in this loop,
|
|
|
|
let's not bother saving on pthread_cond_broadcast().
|
|
|
|
*/
|
|
|
|
if (ticket->m_lock == lock)
|
|
|
|
release_lock(ticket);
|
2009-11-30 23:39:13 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
2009-12-04 00:34:19 +01:00
|
|
|
Downgrade an exclusive lock to shared metadata lock.
|
2009-11-30 23:39:13 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
@param type Type of lock to which exclusive lock should be downgraded.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
*/
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
void MDL_ticket::downgrade_exclusive_lock(enum_mdl_type type)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
2010-02-02 00:22:16 +01:00
|
|
|
mysql_mutex_assert_not_owner(&LOCK_open);
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
/*
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
Do nothing if already downgraded. Used when we FLUSH TABLE under
|
|
|
|
LOCK TABLES and a table is listed twice in LOCK TABLES list.
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
*/
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
if (m_type != MDL_EXCLUSIVE)
|
|
|
|
return;
|
Patch that changes metadata locking subsystem to use mutex per lock and
condition variable per context instead of one mutex and one conditional
variable for the whole subsystem.
This should increase concurrency in this subsystem.
It also opens the way for further changes which are necessary to solve
such bugs as bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock"
and bug #37346 "innodb does not detect deadlock between update and alter
table".
Two other notable changes done by this patch:
- MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object metadata lock is acquired. Now this has
to be done by explicit calls outside of MDL subsystem.
- Instead of using separate MDL_context for opening system tables/tables
for purposes of I_S we now create MDL savepoint in the main context
before opening tables and rollback to this savepoint after closing
them. This means that it is now possible to get ER_LOCK_DEADLOCK error
even not inside a transaction. This might happen in unlikely case when
one runs DDL on one of system tables while also running DDL on some
other tables. Cases when this ER_LOCK_DEADLOCK error is not justified
will be addressed by advanced deadlock detector for MDL subsystem which
we plan to implement.
mysql-test/include/handler.inc:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_innodb.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/handler_myisam.result:
Adjusted handler_myisam.test and handler_innodb.test to the fact that
exclusive metadata locks on tables are now acquired according to
alphabetical order of fully qualified table names instead of order
in which tables are mentioned in statement.
mysql-test/r/mdl_sync.result:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
mysql-test/t/mdl_sync.test:
Adjusted mdl_sync.test to the fact that exclusive metadata locks on
tables are now acquired according to alphabetical order of fully
qualified table names instead of order in which tables are mentioned
in statement.
sql/events.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/ha_ndbcluster.cc:
Since manipulations with open table state no longer install proxy
MDL_context it does not make sense to perform them in order to
satisfy assert in mysql_rm_tables_part2(). Removed them per agreement
with Cluster team. This has not broken test suite since scenario in
which deadlock can occur and assertion fails is not covered by tests.
sql/lock.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/log.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/mdl.cc:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
In fact global lock are now ordinary metadata locks in new GLOBAL
namespace.
To implement these changes:
- Removed LOCK_mdl mutex and COND_mdl condition variable.
- Introduced MDL_lock::m_mutex mutexes which protect individual lock
objects.
- Replaced mdl_locks hash with MDL_map class, which has hash for
MDL_lock objects as a member and separate mutex which protects this
hash. Methods of this class allow to find(), find_or_create() or
remove() MDL_lock objects in concurrency-friendly fashion (i.e.
for most common operation, find_or_create(), we don't acquire
MDL_lock::m_mutex while holding MDL_map::m_mutex. Thanks to MikaelR
for this idea and benchmarks!). Added three auxiliary members to
MDL_lock class (m_is_destroyed, m_ref_usage, m_ref_release) to
support this concurrency-friendly behavior.
- Introduced MDL_context::m_ctx_wakeup_cond condition variable to be
used for waiting until this context's pending request can be
satisfied or its thread has to perform actions to resolve potential
deadlock. Context which want to wait add ticket corresponding to the
request to an appropriate queue of waiters in MDL_lock object so
they can be noticed when other contexts change state of lock and be
awaken by them by signalling on MDL_context::m_ctx_wakeup_cond.
As consequence MDL_ticket objects has to be used for any waiting
in metadata locking subsystem including one which happens in
MDL_context::wait_for_locks() method.
Another consequence is that MDL_context is no longer copyable and
can't be saved/restored when working with system tables.
- Made MDL_lock an abstract class, which delegates specifying exact
compatibility matrix to its descendants. Added MDL_global_lock child
class for global lock (The old is_lock_type_compatible() method
became can_grant_lock() method of this class). Added MDL_object_lock
class to represent per-object lock (The old MDL_lock::can_grant_lock()
became its method). Choice between two classes happens based on MDL
namespace in MDL_lock::create() method.
- Got rid of MDL_lock::type member as its meaning became ambigous for
global locks.
- To simplify waking up of contexts waiting for lock split waiting queue
in MDL_lock class in two queues. One for pending requests for exclusive
(including intention exclusive) locks and another for requests for
shared locks.
- Added virtual wake_up_waiters() method to MDL_lock, MDL_global_lock and
MDL_object_lock classes which allows to wake up waiting contexts after
state of lock changes. Replaced old duplicated code with calls to this
method.
- Adjusted MDL_context::try_acquire_shared_lock()/exclusive_lock()/
global_shared_lock(), MDL_ticket::upgrade_shared_lock_to_exclusive_lock()
and MDL_context::release_ticket() methods to use MDL_map and
MDL_lock::m_mutex instead of single LOCK_mdl mutex and wake up
waiters according to the approach described above. The latter method
also was renamed to MDL_context::release_lock().
- Changed MDL_context::try_acquire_shared_lock()/exclusive_lock() and
release_lock() not to handle global locks. They are now supposed to
be taken explicitly like ordinary metadata locks.
- Added helper MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock() and
acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Since acquiring of several exclusive locks can no longer happen under
single LOCK_mdl mutex the approach to it had to be changed. Now we do
it in one by one fashion. This is done in alphabetical order to avoid
deadlocks. Changed MDL_context::acquire_exclusive_locks() accordingly
(as part of this change moved code responsible for acquiring single
exclusive lock to new MDL_context::acquire_exclusive_lock_impl()
method).
- Since we no longer have single LOCK_mdl mutex which protects all
MDL_context::m_is_waiting_in_mdl members using these members to
determine if we have really awaken context holding conflicting
shared lock became inconvinient. Got rid of this member and changed
notify_shared_lock() helper function and process of acquiring
of/upgrading to exclusive lock not to rely on such information.
Now in MDL_context::acquire_exclusive_lock_impl() and
MDL_ticket::upgrade_shared_lock_to_exclusive_lock() we simply
re-try to wake up threads holding conflicting shared locks after
small time out.
- Adjusted MDL_context::can_wait_lead_to_deadlock() and
MDL_ticket::has_pending_conflicting_lock() to use per-lock
mutexes instead of LOCK_mdl. To do this introduced
MDL_lock::has_pending_exclusive_lock() method.
sql/mdl.h:
Changed metadata locking subsystem to use mutex per lock and condition
variable per context instead of one mutex and one conditional variable
for the whole subsystem. In order to implement this change:
- Added MDL_key::cmp() method to be able to sort MDL_key objects
alphabetically. Changed length fields in MDL_key class to uint16
as 16-bit is enough for length of any key.
- Changed MDL_ticket::get_ctx() to return pointer to non-const
object in order to be able to use MDL_context::awake() method
for such contexts.
- Got rid of unlocked versions of can_wait_lead_to_deadlock()/
has_pending_conflicting_lock() methods in MDL_context and
MDL_ticket. We no longer has single mutex which protects all
locks. Thus one always has to use versions of these methods
which acquire per-lock mutexes.
- MDL_request_list type of list now counts its elements.
- Added MDL_context::m_ctx_wakeup_cond condition variable to be used
for waiting until this context's pending request can be satisfied
or its thread has to perform actions to resolve potential deadlock.
Added awake() method to wake up context from such wait.
Addition of condition variable made MDL_context uncopyable.
As result we no longer can save/restore MDL_context when working
with system tables. Instead we create MDL savepoint before opening
those tables and rollback to it once they are closed.
- MDL_context::release_ticket() became release_lock() method.
- Added auxiliary MDL_context::acquire_exclusive_lock_impl() method
which does all necessary work to acquire exclusive lock on one object
but should not be used directly as it does not enforce any asserts
ensuring that no deadlocks are possible.
- Since we no longer need to know if thread trying to acquire exclusive
lock managed to wake up any threads having conflicting shared locks
(as, anyway, we will try to wake up such threads again shortly)
- MDL_context::m_is_waiting_in_mdl member became unnecessary and
notify_shared_lock() no longer needs to be friend of MDL_context.
Changed approach to handling of global metadata locks. Instead of
implicitly acquiring intention exclusive locks when user requests
per-object upgradeable or exclusive locks now we require them to be
acquired explicitly in the same way as ordinary metadata locks.
- Added new GLOBAL namespace for such locks.
- Added new type of lock to be requested MDL_INTENTION_EXCLISIVE.
- Added MDL_context::try_acquire_global_intention_exclusive_lock()
and acquire_global_intention_exclusive_lock() methods.
- Moved common code from MDL_context::acquire_global_shared_lock()
and acquire_global_intention_exclusive_lock() to new method -
MDL_context::acquire_lock_impl().
- Moved common code from MDL_context::try_acquire_shared_lock(),
try_acquire_global_intention_exclusive_lock()/exclusive_lock()
to MDL_context::try_acquire_lock_impl().
- Added helper MDL_context::is_global_lock_owner() method to be
able easily to find what kind of global lock this context holds.
- MDL_context::m_has_global_shared_lock became unnecessary as
global read lock is now represented by ordinary ticket.
- Removed assert in MDL_context::set_lt_or_ha_sentinel() which became
false for cases when we execute LOCK TABLES under global read lock
mode.
sql/mysql_priv.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result calls
opening/closing system tables were changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sp.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_base.cc:
close_thread_tables():
Since we no longer use separate MDL_context for opening system
tables we need to avoid releasing all transaction locks when
closing system table. Releasing metadata lock on system table
is now responsibility of THD::restore_backup_open_tables_state().
open_table_get_mdl_lock(),
Open_table_context::recover_from_failed_open():
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object upgradable or exclusive metadata lock is
acquired. So this have to be done explicitly from these calls.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
open_table():
Do not release metadata lock if we have failed to open table as
this lock might have been acquired by one of previous statements
in transaction, and therefore should not be released.
open_system_tables_for_read()/close_system_tables()/
open_performance_schema_table():
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
close_performance_schema_table():
Got rid of duplicated code.
sql/sql_class.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. Also releasing
metadata lock on system table is now responsibility of
THD::restore_backup_open_tables_state().
Adjusted assert in THD::cleanup() to take into account fact that now
we also use MDL sentinel for global read lock.
sql/sql_class.h:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. As result:
- 'mdl_context' member was moved out of Open_tables_state to THD class.
enter_locked_tables_mode()/leave_locked_tables_mode() had to follow.
- Methods of THD responsible for saving/restoring open table state were
changed to use Open_tables_backup class which in addition to
Open_table_state has a member for this savepoint.
Changed Open_table_context class to store MDL_request object for
global intention exclusive lock acquired when opening tables.
sql/sql_delete.cc:
MDL subsystem no longer implicitly acquires global intention exclusive
metadata lock when per-object exclusive metadata lock is acquired.
Now this has to be done by explicit calls outside of MDL subsystem.
sql/sql_help.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
sql/sql_parse.cc:
Adjusted assert reload_acl_and_cache() to the fact that global read
lock now takes full-blown metadata lock.
sql/sql_plist.h:
Added support for element counting to I_P_List list template.
One can use policy classes to specify if such counting is needed
or not needed for particular list.
sql/sql_show.cc:
Instead of using separate MDL_context for opening tables for I_S
purposes we now create MDL savepoint in the main context before
opening tables and rollback to this savepoint after closing them.
To support this and similar change for system tables methods of
THD responsible for saving/restoring open table state were changed
to use Open_tables_backup class which in addition to Open_table_state
has a member for this savepoint. As result code opening/closing tables
for I_S purposes was changed to use Open_tables_backup instead of
Open_table_state class as well.
sql/sql_table.cc:
mysql_rm_tables_part2():
Since now global intention exclusive metadata lock is ordinary
metadata lock we no longer can rely that by releasing MDL locks
on all tables we will release all locks acquired by this routine.
So in non-LOCK-TABLES mode we have to release all locks acquired
explicitly.
prepare_for_repair(), mysql_alter_table():
MDL subsystem no longer implicitly acquires global intention
exclusive metadata lock when per-object exclusive metadata lock
is acquired. Now this has to be done by explicit calls outside of
MDL subsystem.
sql/tztime.cc:
Instead of using separate MDL_context for opening system tables we now
create MDL savepoint in the main context before opening such tables
and rollback to this savepoint after closing them. To support this
change methods of THD responsible for saving/restoring open table
state were changed to use Open_tables_backup class which in addition
to Open_table_state has a member for this savepoint. As result code
opening/closing system tables was changed to use Open_tables_backup
instead of Open_table_state class as well.
Also changed code not to use special mechanism for open system tables
when it is not really necessary.
2010-01-21 21:43:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
rw_wrlock(&m_lock->m_rwlock);
|
|
|
|
/*
|
|
|
|
To update state of MDL_lock object correctly we need to temporarily
|
|
|
|
exclude ticket from the granted queue and then include it back.
|
|
|
|
*/
|
|
|
|
m_lock->m_granted.remove_ticket(this);
|
|
|
|
m_type= type;
|
|
|
|
m_lock->m_granted.add_ticket(this);
|
|
|
|
m_lock->wake_up_waiters();
|
|
|
|
rw_unlock(&m_lock->m_rwlock);
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
2009-12-04 00:34:19 +01:00
|
|
|
Auxiliary function which allows to check if we have some kind of lock on
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
a object. Returns TRUE if we have a lock of a given or stronger type.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-09 09:51:20 +01:00
|
|
|
@param mdl_namespace Id of object namespace
|
|
|
|
@param db Name of the database
|
|
|
|
@param name Name of the object
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
@param mdl_type Lock type. Pass in the weakest type to find
|
|
|
|
out if there is at least some lock.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
@return TRUE if current context contains satisfied lock for the object,
|
|
|
|
FALSE otherwise.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
*/
|
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
bool
|
2009-12-10 09:21:38 +01:00
|
|
|
MDL_context::is_lock_owner(MDL_key::enum_mdl_namespace mdl_namespace,
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
const char *db, const char *name,
|
|
|
|
enum_mdl_type mdl_type)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
MDL_request mdl_request;
|
|
|
|
bool is_transactional_unused;
|
|
|
|
mdl_request.init(mdl_namespace, db, name, mdl_type);
|
|
|
|
MDL_ticket *ticket= find_ticket(&mdl_request, &is_transactional_unused);
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
DBUG_ASSERT(ticket == NULL || ticket->m_lock);
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
return ticket;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
Check if we have any pending locks which conflict with existing shared lock.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
@pre The ticket must match an acquired lock.
|
|
|
|
|
2009-12-30 18:53:30 +01:00
|
|
|
@return TRUE if there is a conflicting lock request, FALSE otherwise.
|
|
|
|
*/
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
bool MDL_ticket::has_pending_conflicting_lock() const
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
return m_lock->has_pending_conflicting_lock(m_type);
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
2009-12-04 00:34:19 +01:00
|
|
|
Associate pointer to an opaque object with a lock.
|
|
|
|
|
|
|
|
@param cached_object Pointer to the object
|
|
|
|
@param release_hook Cleanup function to be called when MDL subsystem
|
|
|
|
decides to remove lock or associate another object.
|
|
|
|
|
|
|
|
This is used to cache a pointer to TABLE_SHARE in the lock
|
|
|
|
structure. Such caching can save one acquisition of LOCK_open
|
|
|
|
and one table definition cache lookup for every table.
|
|
|
|
|
|
|
|
Since the pointer may be stored only inside an acquired lock,
|
|
|
|
the caching is only effective when there is more than one lock
|
|
|
|
granted on a given table.
|
|
|
|
|
|
|
|
This function has the following usage pattern:
|
|
|
|
- try to acquire an MDL lock
|
|
|
|
- when done, call for mdl_get_cached_object(). If it returns NULL, our
|
|
|
|
thread has the only lock on this table.
|
|
|
|
- look up TABLE_SHARE in the table definition cache
|
|
|
|
- call mdl_set_cache_object() to assign the share to the opaque pointer.
|
|
|
|
|
|
|
|
The release hook is invoked when the last shared metadata
|
|
|
|
lock on this name is released.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
*/
|
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
void
|
|
|
|
MDL_ticket::set_cached_object(void *cached_object,
|
|
|
|
mdl_cached_object_release_hook release_hook)
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
|
|
|
DBUG_ENTER("mdl_set_cached_object");
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
DBUG_PRINT("enter", ("db=%s name=%s cached_object=%p",
|
2009-12-09 17:11:26 +01:00
|
|
|
m_lock->key.db_name(), m_lock->key.name(),
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
cached_object));
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
/*
|
2009-12-04 00:52:05 +01:00
|
|
|
TODO: This assumption works now since we do get_cached_object()
|
|
|
|
and set_cached_object() in the same critical section. Once
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
this becomes false we will have to call release_hook here and
|
|
|
|
use additional mutex protecting 'cached_object' member.
|
|
|
|
*/
|
2009-12-04 00:52:05 +01:00
|
|
|
DBUG_ASSERT(!m_lock->cached_object);
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
m_lock->cached_object= cached_object;
|
|
|
|
m_lock->cached_object_release_hook= release_hook;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
|
|
|
DBUG_VOID_RETURN;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
2009-12-04 00:34:19 +01:00
|
|
|
Get a pointer to an opaque object that associated with the lock.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
@param ticket Lock ticket for the lock which the object is associated to.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
|
2009-12-04 00:34:19 +01:00
|
|
|
@return Pointer to an opaque object associated with the lock.
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
*/
|
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
void *MDL_ticket::get_cached_object()
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
{
|
2009-12-04 00:52:05 +01:00
|
|
|
return m_lock->cached_object;
|
Initial import of WL#3726 "DDL locking for all metadata objects".
Backport of:
------------------------------------------------------------
revno: 2630.4.1
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Fri 2008-05-23 17:54:03 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
------------------------------------------------------------
This is the first patch in series. It transforms the metadata
locking subsystem to use a dedicated module (mdl.h,cc). No
significant changes in the locking protocol.
The import passes the test suite with the exception of
deprecated/removed 6.0 features, and MERGE tables. The latter
are subject to a fix by WL#4144.
Unfortunately, the original changeset comments got lost in a merge,
thus this import has its own (largely insufficient) comments.
This patch fixes Bug#25144 "replication / binlog with view breaks".
Warning: this patch introduces an incompatible change:
Under LOCK TABLES, it's no longer possible to FLUSH a table that
was not locked for WRITE.
Under LOCK TABLES, it's no longer possible to DROP a table or
VIEW that was not locked for WRITE.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.2
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:03:45 +0400
message:
WL#3726 "DDL locking for all metadata objects".
After review fixes in progress.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.3
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 14:08:51 +0400
message:
WL#3726 "DDL locking for all metadata objects"
Fixed failing Windows builds by adding mdl.cc to the lists
of files needed to build server/libmysqld on Windows.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.4
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sat 2008-05-24 21:57:58 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Fix for assert failures in kill.test which occured when one
tried to kill ALTER TABLE statement on merge table while it
was waiting in wait_while_table_is_used() for other connections
to close this table.
These assert failures stemmed from the fact that cleanup code
in this case assumed that temporary table representing new
version of table was open with adding to THD::temporary_tables
list while code which were opening this temporary table wasn't
always fulfilling this.
This patch changes code that opens new version of table to
always do this linking in. It also streamlines cleanup process
for cases when error occurs while we have new version of table
open.
******
WL#3726 "DDL locking for all metadata objects"
Add libmysqld/mdl.cc to .bzrignore.
******
Backport of:
------------------------------------------------------------
revno: 2630.4.6
committer: Dmitry Lenev <dlenev@mysql.com>
branch nick: mysql-6.0-3726-w
timestamp: Sun 2008-05-25 00:33:22 +0400
message:
WL#3726 "DDL locking for all metadata objects".
Addition to the fix of assert failures in kill.test caused by
changes for this worklog.
Make sure we close the new table only once.
.bzrignore:
Add libmysqld/mdl.cc
libmysqld/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of libmysqld.
libmysqld/Makefile.am:
Added files implementing new meta-data locking subsystem to the server.
mysql-test/include/handler.inc:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/create.result:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/r/flush.result:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/r/flush_table.result:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/r/handler_innodb.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/handler_myisam.result:
Use separate connection for waiting while threads performing DDL
operations conflicting with open HANDLER tables reach blocked
state. This is required because now we check and close tables open
by HANDLER statements in this connection conflicting with DDL in
another each time open_tables() is called and thus select from I_S
which is used for waiting will unblock DDL operations if issued
from connection with open HANDLERs.
mysql-test/r/information_schema.result:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange tests to match 6.0 better (fewer merge conflicts).
mysql-test/r/kill.result:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/r/lock.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/r/partition_column_prune.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/partition_pruning.result:
Update results (same results in 6.0), WL#3726
mysql-test/r/ps_ddl.result:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/r/sp.result:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/r/view.result:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock
mysql-test/r/view_grant.result:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/r/view_multi.result:
Added test case for bug#25144 "replication / binlog with view
breaks".
mysql-test/suite/rpl/t/disabled.def:
Disable test for deprecated features (they don't work with new MDL).
mysql-test/t/create.test:
Adjusted test case after change in implementation of CREATE TABLE
... SELECT. We no longer have special check in open_table() which
catches the case when we select from the table created. Instead we
rely on unique_table() call which happens after opening and
locking all tables.
mysql-test/t/disabled.def:
Disable merge.test, subject of WL#4144
mysql-test/t/flush.test:
FLUSH TABLES WITH READ LOCK can no longer happen under LOCK
TABLES. Updated test accordingly.
mysql-test/t/flush_table.test:
Under LOCK TABLES we no longer allow to do FLUSH TABLES for tables
locked for read. Updated test accordingly.
mysql-test/t/information_schema.test:
Additional test for WL#3726 "DDL locking for all metadata
objects". Check that we use high-priority metadata lock requests
when filling I_S tables.
Rearrange the results for easier merges with 6.0.
mysql-test/t/kill.test:
Added tests checking that DDL and DML statements waiting for
metadata locks can be interrupted by KILL command.
mysql-test/t/lock.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation write locks on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/lock_multi.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/ps_ddl.test:
We no longer invalidate prepared CREATE TABLE ... SELECT statement
if target table changes. This is OK since it is not strictly
necessary.
The first change is wrong, is caused by FLUSH TABLE
now flushing all unused tables. This is a regression that
Dmitri fixed in 6.0 in a follow up patch.
mysql-test/t/sp.test:
Under LOCK TABLES we no longer allow accessing views which were
not explicitly locked. To access view we need to obtain metadata
lock on it and doing this under LOCK TABLES may lead to deadlocks.
mysql-test/t/trigger_notembedded.test:
Adjusted test case to the changes of status in various places
caused by change in implementation FLUSH TABLES WITH READ LOCK,
which is now takes global metadata lock before flushing tables and
therefore waits on at these places.
mysql-test/t/view.test:
One no longer is allowed to do DROP VIEW under LOCK TABLES even if
this view is locked by LOCK TABLES. The problem is that in such
situation even "write locks" on view are not mutually exclusive so
upgrading metadata lock which is required for dropping of view
will lead to deadlock.
mysql-test/t/view_grant.test:
ALTER VIEW implementation was changed to open a view only after
checking that user which does alter has appropriate privileges on
it. This means that in case when user's privileges are
insufficient for this we won't check that new view definer is the
same as original one or user performing alter has SUPER privilege.
Adjusted test case accordingly.
mysql-test/t/view_multi.test:
Added test case for bug#25144 "replication / binlog with view
breaks".
sql/CMakeLists.txt:
Added mdl.cc to the list of files needed for building of server.
sql/Makefile.am:
Added files implementing new meta-data locking subsystem to the
server.
sql/event_db_repository.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when TABLE_LIST objects is also allocated
there or on stack.
sql/ha_ndbcluster.cc:
Adjusted code to work nicely with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/ha_ndbcluster_binlog.cc:
Adjusted code to work with new metadata locking subsystem.
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/handler.cc:
update_frm_version():
Directly update TABLE_SHARE::mysql_version member instead of
going through all TABLE instances for this table (old code was a
legacy from pre-table-definition-cache days).
sql/lock.cc:
Use new metadata locking subsystem. Threw away most of functions
related to name locking as now one is supposed to use metadata
locking API instead. In lock_global_read_lock() and
unlock_global_read_lock() in order to avoid problems with global
read lock sneaking in at the moment when we perform FLUSH TABLES
or ALTER TABLE under LOCK TABLES and when tables being reopened
are protected only by metadata locks we also have to take global
shared meta data lock.
sql/log_event.cc:
Adjusted code to work with new metadata locking subsystem. For
tables open by slave thread for applying RBR events allocate
memory for lock request object in the same chunk of memory as
TABLE_LIST objects for them. In order to ensure that we keep these
objects around until tables are open always close tables before
calling Relay_log_info::clear_tables_to_lock(). Use new auxiliary
Relay_log_info::slave_close_thread_tables() method to enforce
this.
sql/log_event_old.cc:
Adjusted code to work with new metadata locking subsystem. Since
for tables open by slave thread for applying RBR events memory for
lock request object is allocated in the same chunk of memory as
TABLE_LIST objects for them we have to ensure that we keep these
objects around until tables are open. To ensure this we always
close tables before calling
Relay_log_info::clear_tables_to_lock(). To enfore this we use
new auxiliary Relay_log_info::slave_close_thread_tables()
method.
sql/mdl.cc:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mdl.h:
Implemented new metadata locking subsystem and API described in
WL3726 "DDL locking for all metadata objects".
sql/mysql_priv.h:
- close_thread_tables()/close_tables_for_reopen() now has one more
argument which indicates that metadata locks should be released
but not removed from the context in order to be used later in
mdl_wait_for_locks() and tdc_wait_for_old_version().
- close_cached_table() routine is no longer public.
- Thread waiting in wait_while_table_is_used() can be now killed
so this function returns boolean to make caller aware of such
situation.
- We no longer have table cache as separate entity instead used
and unused TABLE instances are linked to TABLE_SHARE objects in
table definition cache.
- Now third argument of open_table() is also used for requesting
table repair or auto-discovery of table's new definition. So its
type was changed from bool to enum.
- Added tdc_open_view() function for opening view by getting its
definition from disk (and table cache in future).
- reopen_name_locked_table() no longer needs "link_in" argument as
now we have exclusive metadata locks instead of dummy TABLE
instances when this function is called.
- find_locked_table() now takes head of list of TABLE instances
instead of always scanning through THD::open_tables list. Also
added find_write_locked_table() auxiliary.
- reopen_tables(), close_cached_tables() no longer have
mark_share_as_old and wait_for_placeholder arguments. Instead of
relying on this parameters and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock.
- We no longer need drop_locked_tables() and
abort_locked_tables().
- mysql_ha_rm_tables() now always assume that LOCK_open is not
acquired by caller.
- Added notify_thread_having_shared_lock() callback invoked by
metadata locking subsystem when acquiring an exclusive lock, for
each thread that has a conflicting shared metadata lock.
- Introduced expel_table_from_cache() as replacement for
remove_table_from_cache() (the main difference is that this new
function assumes that caller follows metadata locking protocol
and never waits).
- Threw away most of functions related to name locking. One should
use new metadata locking subsystem and API instead.
sql/mysqld.cc:
Got rid of call initializing/deinitializing table cache since now
it is embedded into table definition cache. Added calls for
initializing/ deinitializing metadata locking subsystem.
sql/rpl_rli.cc:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/rpl_rli.h:
Introduced auxiliary Relay_log_info::slave_close_thread_tables()
method which is used for enforcing that we always close tables
open for RBR before deallocating TABLE_LIST elements and MDL_LOCK
objects for them.
sql/set_var.cc:
close_cached_tables() no longer has wait_for_placeholder argument.
Instead of relying on this parameter and related behavior FLUSH
TABLES WITH READ LOCK now takes global shared metadata lock.
sql/sp_head.cc:
For tables added to the statement's table list by prelocking
algorithm we allocate these objects either on the same memory as
corresponding table list elements or on THD::locked_tables_root
(if we are building table list for LOCK TABLES).
sql/sql_acl.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_base.cc:
Changed code to use new MDL subsystem. Got rid of separate table
cache. Now used and unused TABLE instances are linked to the
TABLE_SHAREs in table definition cache.
check_unused():
Adjusted code to the fact that we no longer have separate table
cache. Removed dead code.
table_def_free():
Free TABLE instances referenced from TABLE_SHARE objects before
destroying table definition cache.
get_table_share():
Added assert which ensures that noone will be able to access
table (and its share) without acquiring some kind of metadata
lock first.
close_handle_and_leave_table_as_lock():
Adjusted code to the fact that TABLE instances now are linked to
list in TABLE_SHARE.
list_open_tables():
Changed this function to use table definition cache instead of
table cache.
free_cache_entry():
Unlink freed TABLE elements from the list of all TABLE instances
for the table in TABLE_SHARE.
kill_delayed_thread_for_table():
Added auxiliary for killing delayed insert threads for
particular table.
close_cached_tables():
Got rid of wait_for_refresh argument as we now rely on global
shared metadata lock to prevent FLUSH WITH READ LOCK sneaking in
when we are reopening tables. Heavily reworked this function to
use new MDL code and not to rely on separate table cache entity.
close_open_tables():
We no longer have separate table cache.
close_thread_tables():
Release metadata locks after closing all tables. Added skip_mdl
argument which allows us not to remove metadata lock requests
from the context in case when we are going to use this requests
later in mdl_wait_for_locks() and tdc_wait_for_old_versions().
close_thread_table()/close_table_for_reopen():
Since we no longer have separate table cache and all TABLE
instances are linked to TABLE_SHARE objects in table definition
cache we have to link/unlink TABLE object to/from appropriate
lists in the share.
name_lock_locked_table():
Moved redundant code to find_write_locked_table() function and
adjusted code to the fact that wait_while_table_is_used() can
now return with an error if our thread is killed.
reopen_table_entry():
We no longer need "link_in" argument as with MDL we no longer
call this function with dummy TABLE object pre-allocated and
added to the THD::open_tables. Also now we add newly-open TABLE
instance to the list of share's used TABLE instances.
table_cache_insert_placeholder():
Got rid of name-locking legacy.
lock_table_name_if_not_cached():
Moved to sql_table.cc the only place where it is used. It was
also reimplemented using new MDL API.
open_table():
- Reworked this function to use new MDL subsystem.
- Changed code to deal with table definition cache directly
instead of going through separate table cache.
- Now third argument is also used for requesting table repair
or auto-discovery of table's new definition. So its type was
changed from bool to enum.
find_locked_table()/find_write_locked_table():
Accept head of list of TABLE objects as first argument and use
this list instead of always searching in THD::open_tables list.
Also added auxiliary for finding write-locked locked tables.
reopen_table():
Adjusted function to work with new MDL subsystem and to properly
manuipulate with lists of used/unused TABLE instaces in
TABLE_SHARE.
reopen_tables():
Removed mark_share_as_old parameter. Instead of relying on it
and related behavior FLUSH TABLES WITH READ LOCK now takes
global shared metadata lock. Changed code after removing
separate table cache.
drop_locked_tables()/abort_locked_tables():
Got rid of functions which are no longer needed.
unlock_locked_tables():
Moved this function from sql_parse.cc and changed it to release
memory which was used for allocating metadata lock requests for
tables open and locked by LOCK TABLES.
tdc_open_view():
Intoduced function for opening a view by getting its definition
from disk (and table cache in future).
reopen_table_entry():
Introduced function for opening table definitions while holding
exclusive metatadata lock on it.
open_unireg_entry():
Got rid of this function. Most of its functionality is relocated
to open_table() and open_table_fini() functions, and some of it
to reopen_table_entry() and tdc_open_view(). Also code
resposible for auto-repair and auto-discovery of tables was
moved to separate function.
open_table_entry_fini():
Introduced function which contains common actions which finalize
process of TABLE object creation.
auto_repair_table():
Moved code responsible for auto-repair of table being opened
here.
handle_failed_open_table_attempt()
Moved code responsible for handling failing attempt to open
table to one place (retry due to lock conflict/old version,
auto-discovery and repair).
open_tables():
- Flush open HANDLER tables if they have old version of if there
is conflicting metadata lock against them (before this moment
we had this code in open_table()).
- When we open view which should be processed via derived table
on the second execution of prepared statement or stored
routine we still should call open_table() for it in order to
obtain metadata lock on it and prepare its security context.
- In cases when we discover that some special handling of
failure to open table is needed call
handle_failed_open_table_attempt() which handles all such
scenarios.
open_ltable():
Handling of various special scenarios of failure to open a table
was moved to separate handle_failed_open_table_attempt()
function.
remove_db_from_cache():
Removed this function as it is no longer used.
notify_thread_having_shared_lock():
Added callback which is invoked by MDL subsystem when acquiring
an exclusive lock, for each thread that has a conflicting shared
metadata lock.
expel_table_from_cache():
Introduced function for removing unused TABLE instances. Unlike
remove_table_from_cache() it relies on caller following MDL
protocol and having appropriate locks when calling it and thus
does not do any waiting if table is still in use.
tdc_wait_for_old_version():
Added function which allows open_tables() to wait in cases when
we discover that we should back-off due to presence of old
version of table.
abort_and_upgrade_lock():
Use new MDL calls.
mysql_wait_completed_table():
Got rid of unused function.
open_system_tables_for_read/for_update()/performance_schema_table():
Allocate MDL_LOCK objects on execution memory root in cases when
TABLE_LIST objects for corresponding tables is allocated on
stack.
close_performance_schema_table():
Release metadata locks after closing tables.
******
Use I_P_List for free/used tables list in the table share.
sql/sql_binlog.cc:
Use Relay_log_info::slave_close_thread_tables() method to enforce
that we always close tables open for RBR before deallocating
TABLE_LIST elements and MDL_LOCK objects for them.
sql/sql_class.cc:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
sql/sql_class.h:
Added meta-data locking contexts as part of Open_tables_state
context. Also introduced THD::locked_tables_root memory root
which is to be used for allocating MDL_LOCK objects for tables in
LOCK TABLES statement (end of lifetime for such objects is UNLOCK
TABLES so we can't use statement or execution root for them).
Note: handler_mdl_context and locked_tables_root and
mdl_el_root will be removed by subsequent patches.
sql/sql_db.cc:
mysql_rm_db() does not really need to call remove_db_from_cache()
as it drops each table in the database using
mysql_rm_table_part2(), which performs all necessary operations on
table (definition) cache.
sql/sql_delete.cc:
Use the new metadata locking API for TRUNCATE.
sql/sql_handler.cc:
Changed HANDLER implementation to use new metadata locking
subsystem. Note that MDL_LOCK objects for HANDLER tables are
allocated in the same chunk of heap memory as TABLE_LIST object
for those tables.
sql/sql_insert.cc:
mysql_insert():
find_locked_table() now takes head of list of TABLE object as
its argument instead of always scanning through THD::open_tables
list.
handle_delayed_insert():
Allocate metadata lock request object for table open by delayed
insert thread on execution memroot. create_table_from_items():
We no longer allocate dummy TABLE objects for tables being
created if they don't exist. As consequence
reopen_name_locked_table() no longer has link_in argument.
open_table() now has one more argument which is not relevant for
temporary tables.
sql/sql_parse.cc:
- Moved unlock_locked_tables() routine to sql_base.cc and made
available it in other files. Got rid of some redundant code by
using this function.
- Replaced boolean TABLE_LIST::create member with enum
open_table_type member.
- Use special memory root for allocating MDL_LOCK objects for
tables open and locked by LOCK TABLES (these object should live
till UNLOCK TABLES so we can't allocate them on statement nor
execution memory root). Also properly set metadata lock
upgradability attribure for those tables.
- Under LOCK TABLES it is no longer allowed to flush tables which
are not write-locked as this breaks metadata locking protocol
and thus potentially might lead to deadlock.
- Added auxiliary adjust_mdl_locks_upgradability() function.
sql/sql_partition.cc:
Adjusted code to the fact that reopen_tables() no longer has
"mark_share_as_old" argument. Got rid of comments which are no
longer true.
sql/sql_plist.h:
Added I_P_List template class for parametrized intrusive doubly
linked lists and I_P_List_iterator for corresponding iterator.
Unlike for I_List<> list elements of such list can participate in
several lists. Unlike List<> such lists are doubly-linked and
intrusive.
sql/sql_plugin.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_prepare.cc:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take exclusive
metadata lock on it.
sql/sql_rename.cc:
Use new metadata locking subsystem in implementation of RENAME
TABLE.
sql/sql_servers.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables. Got rid of redundant code by using unlock_locked_tables()
function.
sql/sql_show.cc:
Acquire shared metadata lock when we are getting information for
I_S table directly from TABLE_SHARE without doing full-blown table
open. We use high priority lock request in this situation in
order to avoid deadlocks.
Also allocate metadata lock requests objects (MDL_LOCK) on
execution memory root in cases when TABLE_LIST objects are also
allocated there
sql/sql_table.cc:
mysql_rm_table():
Removed comment which is no longer relevant.
mysql_rm_table_part2():
Now caller of mysql_ha_rm_tables() should not own LOCK_open.
Adjusted code to use new metadata locking subsystem instead of
name-locks.
lock_table_name_if_not_cached():
Moved this function from sql_base.cc to this file and
reimplemented it using metadata locking API.
mysql_create_table():
Adjusted code to use new MDL API.
wait_while_table_is_used():
Changed function to use new MDL subsystem. Made thread waiting
in it killable (this also led to introduction of return value so
caller can distinguish successful executions from situations
when waiting was aborted).
close_cached_tables():
Thread waiting in this function is killable now. As result it
has return value for distinguishing between succes and failure.
Got rid of redundant boradcast_refresh() call.
prepare_for_repair():
Use MDL subsystem instead of name-locks.
mysql_admin_table():
mysql_ha_rm_tables() now always assumes that caller doesn't own
LOCK_open.
mysql_repair_table():
We should mark all elements of table list as requiring
upgradable metadata locks.
mysql_create_table_like():
Use new MDL subsystem instead of name-locks.
create_temporary_tables():
We don't need to obtain metadata locks when creating temporary
table.
mysql_fast_or_online_alter_table():
Thread waiting in wait_while_table_is_used() is now killable.
mysql_alter_table():
Adjusted code to work with new MDL subsystem and to the fact
that threads waiting in what_while_table_is_used() and
close_cached_table() are now killable.
sql/sql_test.cc:
We no longer have separate table cache. TABLE instances are now
associated with/linked to TABLE_SHARE objects in table definition
cache.
sql/sql_trigger.cc:
Adjusted code to work with new metadata locking subsystem. Also
reopen_tables() no longer has mark_share_as_old argument (Instead
of relying on this parameter and related behavior FLUSH TABLES
WITH READ LOCK now takes global shared metadata lock).
sql/sql_udf.cc:
Allocate metadata lock requests objects (MDL_LOCK) on execution
memory root in cases when we use stack TABLE_LIST objects to open
tables.
sql/sql_update.cc:
Adjusted code to work with new meta-data locking subsystem.
sql/sql_view.cc:
Added proper meta-data locking to implementations of
CREATE/ALTER/DROP VIEW statements. Now we obtain exclusive
meta-data lock on a view before creating/ changing/dropping it.
This ensures that all concurrent statements that use this view
will finish before our statement will proceed and therefore we
will get correct order of statements in the binary log.
Also ensure that TABLE_LIST::mdl_upgradable attribute is properly
propagated for underlying tables of view.
sql/table.cc:
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock request objects for all elements of table list.
sql/table.h:
TABLE_SHARE:
Got rid of unused members. Introduced members for storing lists
of used and unused TABLE objects for this share.
TABLE:
Added members for linking TABLE objects into per-share lists of
used and unused TABLE instances. Added member for holding
pointer to metadata lock for this table.
TABLE_LIST:
Replaced boolean TABLE_LIST::create member with enum
open_table_type member. This allows easily handle situation in
which instead of opening the table we want only to take
exclusive meta-data lock on it (we need this in order to handle
ALTER VIEW and CREATE VIEW statements).
Introduced new mdl_upgradable member for marking elements of
table list for which we need to take upgradable shared metadata
lock instead of plain shared metadata lock. Added pointer for
holding pointer to MDL_LOCK for the table.
Added auxiliary alloc_mdl_locks() function for allocating metadata
lock requests objects for all elements of table list. Added
auxiliary set_all_mdl_upgradable() function for marking all
elements in table list as requiring upgradable metadata locks.
storage/myisammrg/ha_myisammrg.cc:
Allocate MDL_LOCK objects for underlying tables of MERGE table.
To be reworked once Ingo pushes his patch for WL4144.
2009-11-30 16:55:03 +01:00
|
|
|
}
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
Releases metadata locks that were acquired after a specific savepoint.
|
|
|
|
|
|
|
|
@note Used to release tickets acquired during a savepoint unit.
|
|
|
|
@note It's safe to iterate and unlock any locks after taken after this
|
|
|
|
savepoint because other statements that take other special locks
|
|
|
|
cause a implicit commit (ie LOCK TABLES).
|
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
@param mdl_savepont The last acquired MDL lock when the
|
|
|
|
savepoint was set.
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
*/
|
|
|
|
|
2009-12-04 00:52:05 +01:00
|
|
|
void MDL_context::rollback_to_savepoint(MDL_ticket *mdl_savepoint)
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
{
|
2009-12-04 00:52:05 +01:00
|
|
|
DBUG_ENTER("MDL_context::rollback_to_savepoint");
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
/* If savepoint is NULL, it is from the start of the transaction. */
|
|
|
|
release_locks_stored_before(mdl_savepoint ?
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
mdl_savepoint : m_trans_sentinel);
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
|
|
|
|
DBUG_VOID_RETURN;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
Release locks acquired by normal statements (SELECT, UPDATE,
|
|
|
|
DELETE, etc) in the course of a transaction. Do not release
|
|
|
|
HANDLER locks, if there are any.
|
|
|
|
|
|
|
|
This method is used at the end of a transaction, in
|
|
|
|
implementation of COMMIT (implicit or explicit) and ROLLBACK.
|
|
|
|
*/
|
|
|
|
|
|
|
|
void MDL_context::release_transactional_locks()
|
|
|
|
{
|
|
|
|
DBUG_ENTER("MDL_context::release_transactional_locks");
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
release_locks_stored_before(m_trans_sentinel);
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
DBUG_VOID_RETURN;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
Does this savepoint have this lock?
|
|
|
|
|
|
|
|
@retval TRUE The ticket is older than the savepoint and
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
is not LT, HA or GLR ticket. Thus it belongs
|
|
|
|
to the savepoint.
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
@retval FALSE The ticket is newer than the savepoint
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
or is an LT, HA or GLR ticket.
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
*/
|
|
|
|
|
|
|
|
bool MDL_context::has_lock(MDL_ticket *mdl_savepoint,
|
|
|
|
MDL_ticket *mdl_ticket)
|
|
|
|
{
|
|
|
|
MDL_ticket *ticket;
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
/* Start from the beginning, most likely mdl_ticket's been just acquired. */
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
MDL_context::Ticket_iterator it(m_tickets);
|
|
|
|
bool found_savepoint= FALSE;
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
while ((ticket= it++) && ticket != m_trans_sentinel)
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
{
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
/*
|
|
|
|
First met the savepoint. The ticket must be
|
|
|
|
somewhere after it.
|
|
|
|
*/
|
2009-12-04 00:52:05 +01:00
|
|
|
if (ticket == mdl_savepoint)
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
found_savepoint= TRUE;
|
|
|
|
/*
|
|
|
|
Met the ticket. If we haven't yet met the savepoint,
|
|
|
|
the ticket is newer than the savepoint.
|
|
|
|
*/
|
|
|
|
if (ticket == mdl_ticket)
|
|
|
|
return found_savepoint;
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
}
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
/* Reached m_trans_sentinel. The ticket must be LT, HA or GRL ticket. */
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
return FALSE;
|
Backport of:
------------------------------------------------------------
revno: 2617.23.18
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 4284-6.0
timestamp: Mon 2009-03-02 18:18:26 -0300
message:
Bug#989: If DROP TABLE while there's an active transaction, wrong binlog order
WL#4284: Transactional DDL locking
This is a prerequisite patch:
These changes are intended to split lock requests from granted
locks and to allow the memory and lifetime of granted locks to
be managed within the MDL subsystem. Furthermore, tickets can
now be shared and therefore are used to satisfy multiple lock
requests, but only shared locks can be recursive.
The problem is that the MDL subsystem morphs lock requests into
granted locks locks but does not manage the memory and lifetime
of lock requests, and hence, does not manage the memory of
granted locks either. This can be problematic because it puts the
burden of tracking references on the users of the subsystem and
it can't be easily done in transactional contexts where the locks
have to be kept around for the duration of a transaction.
Another issue is that recursive locks (when the context trying to
acquire a lock already holds a lock on the same object) requires
that each time the lock is granted, a unique lock request/granted
lock structure structure must be kept around until the lock is
released. This can lead to memory leaks in transactional contexts
as locks taken during the transaction should only be released at
the end of the transaction. This also leads to unnecessary wake
ups (broadcasts) in the MDL subsystem if the context still holds
a equivalent of the lock being released.
These issues are exacerbated due to the fact that WL#4284 low-level
design says that the implementation should "2) Store metadata locks
in transaction memory root, rather than statement memory root" but
this is not possible because a memory root, as implemented in mysys,
requires all objects allocated from it to be freed all at once.
This patch combines review input and significant code contributions
from Konstantin Osipov (kostja) and Dmitri Lenev (dlenev).
mysql-test/r/mdl_sync.result:
Add test case result.
mysql-test/t/mdl_sync.test:
Add test case for shared lock upgrade case.
sql/event_db_repository.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/ha_ndbcluster_binlog.cc:
Use new function names to initialize MDL lock requests.
sql/lock.cc:
Rename MDL functions.
sql/log_event.cc:
The MDL request now holds the table and database name data (MDL_KEY).
sql/mdl.cc:
Move the MDL key to the MDL_LOCK structure in order to make the
object suitable for allocation from a fixed-size allocator. This
allows the simplification of the lists in the MDL_LOCK object,
which now are just two, one for granted tickets and other for
waiting (upgraders) tickets.
Recursive requests for a shared lock on the same object can now
be granted using the same lock ticket. This schema is only used
for shared locks because that the only case that matters. This
is used to avoid waste of resources in case a context (connection)
already holds a shared lock on a object.
sql/mdl.h:
Introduce a metadata lock object key which is used to uniquely
identify lock objects.
Separate the structure used to represent pending lock requests
from the structure used to represent granted metadata locks.
Rename functions used to manipulate locks requests in order to
have a more consistent function naming schema.
sql/sp_head.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/sql_acl.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_base.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_class.h:
Last acquired lock before the savepoint was set.
sql/sql_delete.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_handler.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_insert.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_parse.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_plist.h:
Typedef for iterator type.
sql/sql_plugin.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_servers.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/sql_show.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_table.cc:
Various changes to accommodate that lock requests are separated
from lock tickets (granted locks).
sql/sql_trigger.cc:
Save reference to the lock ticket so it can be downgraded later.
sql/sql_udf.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
sql/table.cc:
Rename mdl_alloc_lock to mdl_request_alloc.
sql/table.h:
Separate MDL lock requests from lock tickets (granted locks).
storage/myisammrg/ha_myisammrg.cc:
Rename alloc_mdl_locks to alloc_mdl_requests.
2009-12-04 00:29:40 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
/**
|
|
|
|
Rearrange the ticket to reside in the part of the list that's
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
beyond m_trans_sentinel. This effectively changes the ticket
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
life cycle, from automatic to manual: i.e. the ticket is no
|
|
|
|
longer released by MDL_context::release_transactional_locks() or
|
|
|
|
MDL_context::rollback_to_savepoint(), it must be released manually.
|
|
|
|
*/
|
|
|
|
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
void MDL_context::move_ticket_after_trans_sentinel(MDL_ticket *mdl_ticket)
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
{
|
|
|
|
m_tickets.remove(mdl_ticket);
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
if (m_trans_sentinel == NULL)
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
{
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
m_trans_sentinel= mdl_ticket;
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
/* sic: linear from the number of transactional tickets acquired so-far! */
|
|
|
|
m_tickets.push_back(mdl_ticket);
|
|
|
|
}
|
|
|
|
else
|
Implement new type-of-operation-aware metadata locks.
Add a wait-for graph based deadlock detector to the
MDL subsystem.
Fixes bug #46272 "MySQL 5.4.4, new MDL: unnecessary deadlock" and
bug #37346 "innodb does not detect deadlock between update and
alter table".
The first bug manifested itself as an unwarranted abort of a
transaction with ER_LOCK_DEADLOCK error by a concurrent ALTER
statement, when this transaction tried to repeat use of a
table, which it has already used in a similar fashion before
ALTER started.
The second bug showed up as a deadlock between table-level
locks and InnoDB row locks, which was "detected" only after
innodb_lock_wait_timeout timeout.
A transaction would start using the table and modify a few
rows.
Then ALTER TABLE would come in, and start copying rows
into a temporary table. Eventually it would stumble on
the modified records and get blocked on a row lock.
The first transaction would try to do more updates, and get
blocked on thr_lock.c lock.
This situation of circular wait would only get resolved
by a timeout.
Both these bugs stemmed from inadequate solutions to the
problem of deadlocks occurring between different
locking subsystems.
In the first case we tried to avoid deadlocks between metadata
locking and table-level locking subsystems, when upgrading shared
metadata lock to exclusive one.
Transactions holding the shared lock on the table and waiting for
some table-level lock used to be aborted too aggressively.
We also allowed ALTER TABLE to start in presence of transactions
that modify the subject table. ALTER TABLE acquires
TL_WRITE_ALLOW_READ lock at start, and that block all writes
against the table (naturally, we don't want any writes to be lost
when switching the old and the new table). TL_WRITE_ALLOW_READ
lock, in turn, would block the started transaction on thr_lock.c
lock, should they do more updates. This, again, lead to the need
to abort such transactions.
The second bug occurred simply because we didn't have any
mechanism to detect deadlocks between the table-level locks
in thr_lock.c and row-level locks in InnoDB, other than
innodb_lock_wait_timeout.
This patch solves both these problems by moving lock conflicts
which are causing these deadlocks into the metadata locking
subsystem, thus making it possible to avoid or detect such
deadlocks inside MDL.
To do this we introduce new type-of-operation-aware metadata
locks, which allow MDL subsystem to know not only the fact that
transaction has used or is going to use some object but also what
kind of operation it has carried out or going to carry out on the
object.
This, along with the addition of a special kind of upgradable
metadata lock, allows ALTER TABLE to wait until all
transactions which has updated the table to go away.
This solves the second issue.
Another special type of upgradable metadata lock is acquired
by LOCK TABLE WRITE. This second lock type allows to solve the
first issue, since abortion of table-level locks in event of
DDL under LOCK TABLES becomes also unnecessary.
Below follows the list of incompatible changes introduced by
this patch:
- From now on, ALTER TABLE and CREATE/DROP TRIGGER SQL (i.e. those
statements that acquire TL_WRITE_ALLOW_READ lock)
wait for all transactions which has *updated* the table to
complete.
- From now on, LOCK TABLES ... WRITE, REPAIR/OPTIMIZE TABLE
(i.e. all statements which acquire TL_WRITE table-level lock) wait
for all transaction which *updated or read* from the table
to complete.
As a consequence, innodb_table_locks=0 option no longer applies
to LOCK TABLES ... WRITE.
- DROP DATABASE, DROP TABLE, RENAME TABLE no longer abort
statements or transactions which use tables being dropped or
renamed, and instead wait for these transactions to complete.
- Since LOCK TABLES WRITE now takes a special metadata lock,
not compatible with with reads or writes against the subject table
and transaction-wide, thr_lock.c deadlock avoidance algorithm
that used to ensure absence of deadlocks between LOCK TABLES
WRITE and other statements is no longer sufficient, even for
MyISAM. The wait-for graph based deadlock detector of MDL
subsystem may sometimes be necessary and is involved. This may
lead to ER_LOCK_DEADLOCK error produced for multi-statement
transactions even if these only use MyISAM:
session 1: session 2:
begin;
update t1 ... lock table t2 write, t1 write;
-- gets a lock on t2, blocks on t1
update t2 ...
(ER_LOCK_DEADLOCK)
- Finally, support of LOW_PRIORITY option for LOCK TABLES ... WRITE
was abandoned.
LOCK TABLE ... LOW_PRIORITY WRITE from now on has the same
priority as the usual LOCK TABLE ... WRITE.
SELECT HIGH PRIORITY no longer trumps LOCK TABLE ... WRITE in
the wait queue.
- We do not take upgradable metadata locks on implicitly
locked tables. So if one has, say, a view v1 that uses
table t1, and issues:
LOCK TABLE v1 WRITE;
FLUSH TABLE t1; -- (or just 'FLUSH TABLES'),
an error is produced.
In order to be able to perform DDL on a table under LOCK TABLES,
the table must be locked explicitly in the LOCK TABLES list.
mysql-test/include/handler.inc:
Adjusted test case to trigger an execution path on which bug 41110
"crash with handler command when used concurrently with alter
table" and bug 41112 "crash in mysql_ha_close_table/get_lock_data
with alter table" were originally discovered. Left old test case
which no longer triggers this execution path for the sake of
coverage.
Added test coverage for HANDLER SQL statements and type-aware
metadata locks.
Added a test for the global shared lock and HANDLER SQL.
Updated tests to take into account that the old simple deadlock
detection heuristics was replaced with a graph-based deadlock
detector.
mysql-test/r/debug_sync.result:
Updated results (see debug_sync.test).
mysql-test/r/handler_innodb.result:
Updated results (see handler.inc test).
mysql-test/r/handler_myisam.result:
Updated results (see handler.inc test).
mysql-test/r/innodb-lock.result:
Updated results (see innodb-lock.test).
mysql-test/r/innodb_mysql_lock.result:
Updated results (see innodb_mysql_lock.test).
mysql-test/r/lock.result:
Updated results (see lock.test).
mysql-test/r/lock_multi.result:
Updated results (see lock_multi.test).
mysql-test/r/lock_sync.result:
Updated results (see lock_sync.test).
mysql-test/r/mdl_sync.result:
Updated results (see mdl_sync.test).
mysql-test/r/sp-threads.result:
SHOW PROCESSLIST output has changed due to the fact that waiting
for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/r/truncate_coverage.result:
Updated results (see truncate_coverage.test).
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
SELECT FROM I_S.PROCESSLIST output has changed due to fact that
waiting for LOCK TABLES WRITE now happens within metadata locking
subsystem.
mysql-test/suite/rpl/t/rpl_sp.test:
Updated to a new SHOW PROCESSLIST state name.
mysql-test/t/debug_sync.test:
Use LOCK TABLES READ instead of LOCK TABLES WRITE as the latter
no longer allows to trigger execution path involving waiting on
thr_lock.c lock and therefore reaching debug sync-point covered
by this test.
mysql-test/t/innodb-lock.test:
Adjusted test case to the fact that innodb_table_locks=0 option is
no longer supported, since LOCK TABLES WRITE handles all its
conflicts within MDL subsystem.
mysql-test/t/innodb_mysql_lock.test:
Added test for bug #37346 "innodb does not detect deadlock between
update and alter table".
mysql-test/t/lock.test:
Added test coverage which checks the fact that we no longer support
DDL under LOCK TABLES on tables which were locked implicitly.
Adjusted existing test cases accordingly.
mysql-test/t/lock_multi.test:
Added test for bug #46272 "MySQL 5.4.4, new MDL: unnecessary
deadlock". Adjusted other test cases to take into account the
fact that waiting for LOCK TABLES ... WRITE now happens within MDL
subsystem.
mysql-test/t/lock_sync.test:
Since LOCK TABLES ... WRITE now takes SNRW metadata lock for
tables locked explicitly we have to implicitly lock InnoDB tables
(through view) to trigger the table-level lock conflict between
TL_WRITE and TL_WRITE_ALLOW_WRITE.
mysql-test/t/mdl_sync.test:
Added basic test coverage for type-of-operation-aware metadata
locks. Also covered with tests some use cases involving HANDLER
statements in which a deadlock could arise.
Adjusted existing tests to take type-of-operation-aware MDL into
account.
mysql-test/t/multi_update.test:
Update to a new SHOW PROCESSLIST state name.
mysql-test/t/truncate_coverage.test:
Adjusted test case after making LOCK TABLES WRITE to wait until
transactions that use the table to be locked are completed.
Updated to the changed name of DEBUG_SYNC point.
sql/handler.cc:
Global read lock functionality has been
moved into a class.
sql/lock.cc:
Global read lock functionality has been
moved into a class.
Updated code to use the new MDL API.
sql/mdl.cc:
Introduced new type-of-operation aware metadata locks.
To do this:
- Changed MDL_lock to use one list for waiting requests and one
list for granted requests. For each list, added a bitmap
that holds information what lock types a list contains.
Added a helper class MDL_lock::List to manipulate with granted
and waited lists while keeping the bitmaps in sync
with list contents.
- Changed lock-compatibility functions to use bitmaps that
define compatibility.
- Introduced a graph based deadlock detector inspired by
waiting_threads.c from Maria implementation.
- Now that we have a deadlock detector, and no longer have
a global lock to protect individual lock objects, but rather
use an rw lock per object, removed redundant code for upgrade,
and the global read lock. Changed the MDL API to
no longer require the caller to acquire the global
intention exclusive lock by means of a separate method.
Removed a few more methods that became redundant.
- Removed deadlock detection heuristic, it has been made
obsolete by the deadlock detector.
- With operation-type-aware metadata locks, MDL subsystem has
become aware of potential conflicts between DDL and open
transactions. This made it possible to remove calls to
mysql_abort_transactions_with_shared_lock() from acquisition
paths for exclusive lock and lock upgrade. Now we can simply
wait for these transactions to complete without fear of
deadlock. Function mysql_lock_abort() has also become
unnecessary for all conflicting cases except when a DDL
conflicts with a connection that has an open HANDLER.
sql/mdl.h:
Introduced new type-of-operation aware metadata locks.
Introduced a graph based deadlock detector and supporting
methods.
Added comments.
God rid of redundant API calls.
Renamed m_lt_or_ha_sentinel to m_trans_sentinel,
since now it guards the global read lock as well as
LOCK TABLES and HANDLER locks.
sql/mysql_priv.h:
Moved the global read lock functionality into a
class.
Added MYSQL_OPEN_FORCE_SHARED_MDL flag which forces
open_tables() to take MDL_SHARED on tables instead of
metadata locks specified in the parser. We use this to
allow PREPARE run concurrently in presence of
LOCK TABLES ... WRITE.
Added signature for find_table_for_mdl_ugprade().
sql/set_var.cc:
Global read lock functionality has been
moved into a class.
sql/sp_head.cc:
When creating TABLE_LIST elements for prelocking or
system tables set the type of request for metadata
lock according to the operation that will be performed
on the table.
sql/sql_base.cc:
- Updated code to use the new MDL API.
- In order to avoid locks starvation we take upgradable
locks all at once. As result implicitly locked tables no
longer get an upgradable lock. Consequently DDL and FLUSH
TABLES for such tables is prohibited.
find_write_locked_table() was replaced by
find_table_for_mdl_upgrade() function.
open_table() was adjusted to return TABLE instance with
upgradable ticket when necessary.
- We no longer wait for all locks on OT_WAIT back off
action -- only on the lock that caused the wait
conflict. Moreover, now we distinguish cases when we
have to wait due to conflict in MDL and old version
of table in TDC.
- Upate mysql_notify_threads_having_share_locks()
to only abort thr_lock.c waits of threads that
have open HANDLERs, since lock conflicts with only
these threads now can lead to deadlocks not detectable
by the MDL deadlock detector.
- Remove mysql_abort_transactions_with_shared_locks()
which is no longer needed.
sql/sql_class.cc:
Global read lock functionality has been moved into a class.
Re-arranged code in THD::cleanup() to simplify assert.
sql/sql_class.h:
Introduced class to incapsulate global read lock
functionality.
Now sentinel in MDL subsystem guards the global read lock
as well as LOCK TABLES and HANDLER locks. Adjusted code
accordingly.
sql/sql_db.cc:
Global read lock functionality has been moved into a class.
sql/sql_delete.cc:
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
TRUNCATE TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_handler.cc:
Inform MDL_context about presence of open HANDLERs.
Since HANLDERs break MDL protocol by acquiring table-level
lock while holding only S metadata lock on a table MDL
subsystem should take special care about such contexts (Now
this is the only case when mysql_lock_abort() is used).
sql/sql_parse.cc:
Global read lock functionality has been moved into a class.
Do not take upgradable metadata locks when opening tables
for CREATE TABLE SELECT as it is not necessary and limits
concurrency.
When initializing TABLE_LIST objects before adding them
to the table list set the type of request for metadata lock
according to the operation that will be performed on the
table.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result FLUSH
TABLES is no longer allowed for such tables.
sql/sql_prepare.cc:
Use MYSQL_OPEN_FORCE_SHARED_MDL flag when opening
tables during PREPARE. This allows PREPARE to run
concurrently in presence of LOCK TABLES ... WRITE.
sql/sql_rename.cc:
Global read lock functionality has been moved into a class.
sql/sql_show.cc:
Updated code to use the new MDL API.
sql/sql_table.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result DROP
TABLE is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_trigger.cc:
Global read lock functionality has been moved into a class.
We no longer acquire upgradable metadata locks on tables
which are locked by LOCK TABLES implicitly. As result
CREATE/DROP TRIGGER is no longer allowed for such tables.
Updated code to use the new MDL API.
sql/sql_view.cc:
Global read lock functionality has been moved into a class.
Fixed results of wrong merge that led to misuse of GLR API.
CREATE VIEW statement is not a commit statement.
sql/table.cc:
When resetting TABLE_LIST objects for PS or SP re-execution
set the type of request for metadata lock according to the
operation that will be performed on the table. Do the same
in auxiliary function initializing metadata lock requests
in a table list.
sql/table.h:
When initializing TABLE_LIST objects set the type of request
for metadata lock according to the operation that will be
performed on the table.
sql/transaction.cc:
Global read lock functionality has been moved into a class.
2010-02-01 12:43:06 +01:00
|
|
|
m_tickets.insert_after(m_trans_sentinel, mdl_ticket);
|
A prerequisite patch for the fix for Bug#46224
"HANDLER statements within a transaction might lead to deadlocks".
Introduce a notion of a sentinel to MDL_context. A sentinel
is a ticket that separates all tickets in the context into two
groups: before and after it. Currently we can have (and need) only
one designated sentinel -- it separates all locks taken by LOCK
TABLE or HANDLER statement, which must survive COMMIT and ROLLBACK
and all other locks, which must be released at COMMIT or ROLLBACK.
The tricky part is maintaining the sentinel up to date when
someone release its corresponding ticket. This can happen, e.g.
if someone issues DROP TABLE under LOCK TABLES (generally,
see all calls to release_all_locks_for_name()).
MDL_context::release_ticket() is modified to take care of it.
******
A fix and a test case for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
An attempt to mix HANDLER SQL statements, which are transaction-
agnostic, an open multi-statement transaction,
and DDL against the involved tables (in a concurrent connection)
could lead to a deadlock. The deadlock would occur when
HANDLER OPEN or HANDLER READ would have to wait on a conflicting
metadata lock. If the connection that issued HANDLER statement
also had other metadata locks (say, acquired in scope of a
transaction), a classical deadlock situation of mutual wait
could occur.
Incompatible change: entering LOCK TABLES mode automatically
closes all open HANDLERs in the current connection.
Incompatible change: previously an attempt to wait on a lock
in a connection that has an open HANDLER statement could wait
indefinitely/deadlock. After this patch, an error ER_LOCK_DEADLOCK
is produced.
The idea of the fix is to merge thd->handler_mdl_context
with the main mdl_context of the connection, used for transactional
locks. This makes deadlock detection possible, since all waits
with locks are "visible" and available to analysis in a single
MDL context of the connection.
Since HANDLER locks and transactional locks have a different life
cycle -- HANDLERs are explicitly open and closed, and so
are HANDLER locks, explicitly acquired and released, whereas
transactional locks "accumulate" till the end of a transaction
and are released only with COMMIT, ROLLBACK and ROLLBACK TO SAVEPOINT,
a concept of "sentinel" was introduced to MDL_context.
All locks, HANDLER and others, reside in the same linked list.
However, a selected element of the list separates locks with
different life cycle. HANDLER locks always reside at the
end of the list, after the sentinel. Transactional locks are
prepended to the beginning of the list, before the sentinel.
Thus, ROLLBACK, COMMIT or ROLLBACK TO SAVEPOINT, only
release those locks that reside before the sentinel. HANDLER locks
must be released explicitly as part of HANDLER CLOSE statement,
or an implicit close.
The same approach with sentinel
is also employed for LOCK TABLES locks. Since HANDLER and LOCK TABLES
statement has never worked together, the implementation is
made simple and only maintains one sentinel, which is used either
for HANDLER locks, or for LOCK TABLES locks.
mysql-test/include/handler.inc:
Add test coverage for Bug#46224 "HANDLER statements within a
transaction might lead to deadlocks".
Extended HANDLER coverage to cover a mix of HANDLER, transactions
and DDL statements.
mysql-test/r/handler_innodb.result:
Update results (Bug#46224).
mysql-test/r/handler_myisam.result:
Update results (Bug#46224).
sql/lock.cc:
Remove thd->some_tables_deleted, it's never used.
sql/log_event.cc:
No need to check for thd->locked_tables_mode,
it's done inside release_transactional_locks().
sql/mdl.cc:
Implement the concept of HANDLER and LOCK TABLES "sentinel".
Implement a method to clone an acquired ticket.
Do not return tickets beyond the sentinel when acquiring
locks, create a copy.
Remove methods to merge and backup MDL_context, they are now
not used (Hurra!). This opens a path to a proper constructor
and destructor of class MDL_context (to be done in a separate
patch).
Modify find_ticket() to provide information about where
the ticket position is with regard to the sentinel.
sql/mdl.h:
Add declarations necessary for the implementation of the concept
of "sentinel", a dedicated ticket separating transactional and
non-transactional locks.
sql/mysql_priv.h:
Add mark_tmp_table_for_reuse() declaration,
a function to "close" a single session (temporary) table.
sql/sql_base.cc:
Remove thd->some_tables_deleted.
Modify deadlock-prevention asserts and deadlock detection
heuristics to take into account that from now on HANDLER locks
reside in the same locking context.
Add broadcast_refresh() to mysql_notify_thread_having_shared_lock():
this is necessary for the case when a thread having a shared lock
is asleep in tdc_wait_for_old_versions(). This situation is only
possible with HANDLER t1 OPEN; FLUSH TABLE (since all over code paths
that lead to tdc_wait_for_old_versions() always have an
empty MDL_context). Previously the server would simply deadlock
in this situation.
sql/sql_class.cc:
Remove now unused member "THD::some_tables_deleted".
Move mysql_ha_cleanup() a few lines above in THD::cleanup()
to make sure that all handlers are closed when it's time to
destroy the MDL_context of this connection.
Remove handler_mdl_context and handler_tables.
sql/sql_class.h:
Remove THD::handler_tables, THD::handler_mdl_context,
THD::some_tables_deleted.
sql/sql_handler.cc:
Remove thd->handler_tables.
Remove thd->handler_mdl_context.
Rewrite mysql_ha_open() to have no special provision for MERGE
tables, now that we don't have to manipulate with thd->handler_tables
it's easy to do.
Remove dead code.
Fix a bug in mysql_ha_flush() when we would always flush
a temporary HANDLER when mysql_ha_flush() is called (actually
mysql_ha_flush() never needs to flush temporary tables).
sql/sql_insert.cc:
Update a comment, no more thd->some_tables_deleted.
sql/sql_parse.cc:
Implement an incompatible change: entering LOCK TABLES closes
active HANDLERs, if any.
Now that we have a sentinel, we don't need to check
for thd->locked_tables_mode when releasing metadata locks in
COMMIT/ROLLBACK.
sql/sql_plist.h:
Add new (now necessary) methods to the list class.
sql/sql_prepare.cc:
Make sure we don't release HANDLER locks when rollback to a
savepoint, set to not keep locks taken at PREPARE.
sql/sql_servers.cc:
Update to a new signature of MDL_context::release_all_locks().
sql/sql_table.cc:
Remove thd->some_tables_deleted.
sql/transaction.cc:
Add comments.
Make sure rollback to (MDL) savepoint works under LOCK TABLES and
with HANDLER tables.
2009-12-22 17:09:15 +01:00
|
|
|
}
|